Scolaris Content Display Scolaris Content Display

Inhibidores selectivos de la recaptación de serotonina (ISRS) para los trastornos del espectro autista (TEA)

Contraer todo Desplegar todo

Resumen

disponible en

Antecedentes

Los trastornos del espectro autista (TEA) se caracterizan por anomalías en la interacción social y la capacidad de comunicación, así como por conductas estereotipadas y actividades e intereses restringidos. Los inhibidores selectivos de la recaptación de serotonina (ISRS) se prescriben para el tratamiento de enfermedades a menudo concomitantes con los TEA como la depresión, la ansiedad y las conductas obsesivo‐compulsivas.

Objetivos

Determinar si el tratamiento con un ISRS:
1. mejora las características fundamentales del autismo (problemas de interacción social, de comunicación y conductuales);
2. mejora los aspectos no fundamentales de la conducta o el funcionamiento, como las lesiones autoprovocadas;
3. mejora la calidad de vida de adultos o niños y sus cuidadores;
4. tiene efectos a corto y a largo plazo sobre los resultados;
5. tiene efectos perjudiciales.

Métodos de búsqueda

Se hicieron búsquedas en las siguientes bases de datos hasta marzo de 2013: CENTRAL, Ovid MEDLINE, Embase, CINAHL, PsycINFO, ERIC y en Sociological Abstracts. También se hicieron búsquedas en ClinicalTrials.gov y en la International Clinical Trials Registry Platform (ICTRP). Lo anterior se complementó con búsquedas en las listas de referencias y mediante contacto con expertos reconocidos en el tema.

Criterios de selección

Ensayos controlados aleatorios (ECA) de cualquier dosis de ISRS oral en comparación con placebo en pacientes con TEA.

Obtención y análisis de los datos

Dos revisores, de forma independiente, seleccionaron los estudios para inclusión, extrajeron los datos y evaluaron el riesgo de sesgo de cada estudio.

Resultados principales

Se incluyeron nueve ECA con un total de 320 participantes. Se evaluaron cuatro ISRS: fluoxetina (tres estudios), fluvoxamina (dos estudios), fenfluramina (dos estudios) y citalopram (dos estudios). Cinco estudios incluyeron solamente a niños y cuatro estudios incluyeron solamente a adultos. Se utilizaron criterios de inclusión variables con respecto a los criterios de diagnóstico y al coeficiente intelectual de los participantes. Se informaron 18 medidas de resultado diferentes. Aunque más de un estudio informó datos de la Clinical Global Impression (CGI) y la conducta obsesivo‐compulsiva (COC), en cada estudio se utilizaron diferentes tipos de herramientas o componentes de estos resultados. Por lo tanto, los datos no fueron apropiados para el metanálisis excepto para un resultado (proporción de mejoría). Un estudio grande de alta calidad realizado en niños no mostró pruebas de un efecto positivo del citalopram. Tres estudios pequeños en adultos mostraron resultados positivos en la CGI y la COC; un estudio mostró mejorías en la agresión y otro en la ansiedad.

Conclusiones de los autores

No existen pruebas de un efecto de los ISRS en los niños y surgieron pruebas de efectos perjudiciales. Hay pruebas limitadas de la efectividad de los ISRS en adultos a partir de estudios pequeños en los que el riesgo de sesgo fue incierto.

PICO

Population
Intervention
Comparison
Outcome

El uso y la enseñanza del modelo PICO están muy extendidos en el ámbito de la atención sanitaria basada en la evidencia para formular preguntas y estrategias de búsqueda y para caracterizar estudios o metanálisis clínicos. PICO son las siglas en inglés de cuatro posibles componentes de una pregunta de investigación: paciente, población o problema; intervención; comparación; desenlace (outcome).

Para saber más sobre el uso del modelo PICO, puede consultar el Manual Cochrane.

Resumen en términos sencillos

disponible en

Inhibidores selectivos de la recaptación de serotonina para el tratamiento de los pacientes con trastornos del espectro autista

Los trastornos del espectro autista (TEA) se caracterizan por problemas con la interacción social y la comunicación, así como por conductas repetitivas y actividades e intereses limitados. Los inhibidores selectivos de la recaptación de serotonina (ISRS) son una clase de antidepresivos que en ocasiones se administran para reducir la ansiedad o las conductas obsesivo‐compulsivas. Se encontraron nueve ensayos con 320 pacientes que evaluaron cuatro ISRS: fluoxetina, fluvoxamina, fenfluramina y citalopram. Cinco estudios incluyeron solamente a niños y cuatro estudios incluyeron solamente a adultos. Un ensayo incluyó a 149 niños, pero los otros ensayos fueron mucho menores. No se encontraron ensayos que evaluaran la sertralina, la paroxetina ni el escitalopram. No existen pruebas que apoyen el uso de los ISRS para tratar el autismo en niños. Hay pruebas limitadas, que todavía no son suficientemente consistentes, que indican la efectividad de los ISRS en adultos con autismo. El tratamiento con un ISRS puede provocar efectos secundarios. Las decisiones acerca del uso de los ISRS en indicaciones clínicas establecidas que pueden concomitar con el autismo, como el trastorno obsesivo‐compulsivo y la depresión en adultos o niños, y la ansiedad en adultos, se deben tomar de forma individual.

Authors' conclusions

Implications for practice

There is no evidence that selective serotonin reuptake inhibitors (SSRIs) are effective as a treatment for children with autism. In fact, there is emerging evidence that they are not effective and can cause harm.

For adults, small positive effects have been seen with fewer side effects reported with fluoxetine and fluvoxamine, but the possible risk of bias and small sample size of the trials mean there is not strong evidence to support these treatments. A small study of citalopram in adults with high levels of repetitive behaviours has shown no positive effects.

Decisions about the use of SSRIs for established clinical indications that may co‐occur with autism, such as obsessive‐compulsive disorder and depression, and anxiety (in the case of adults), should be made on a case‐by‐case basis.

Not all the SSRIs currently in use have been subject to controlled trials for autism spectrum disorders (ASD). As ASD causes substantial impairment, parents of children with the condition are motivated to try treatments regardless of the evidence. Nevertheless, it is important that prescribing clinicians are explicit with parents and patients about the limited evidence, discuss the risks of treatment, and discuss other pharmacological and non‐pharmacological interventions. 

Implications for research

The present review has highlighted the significant challenges in researching outcomes in the pharmacological treatment of autism. However, high‐quality studies are feasible if adequately resourced, as demonstrated by the King 2009 trial of citalopram reported in this review, and the trial of the unrelated compound risperidone (McCracken 2002), presented in another review (Jesner 2007).

In our opinion, knowledge about the effectiveness and safety of SSRIs for childhood autism would be best served in the first instance by a replication of the citalopram study, which will either confirm or refute the absence of effect on core symptoms. For completeness, an adequately‐powered randomised controlled trial (RCT) should be conducted on at least one other SSRI. We would recommend fluoxetine owing to its favourable safety profile. We are aware of one such study that reached primary study completion in 2009 (NCT00515320) and another (ACTRN12608000173392) that began in 2010 and is still recruiting. Another study assessing feasibility and safety in preschool children was due to be completed in 2008 (NCT00183339). Sufficiently large trials would permit the examination of subgroup differences in responsiveness to SSRIs. Comparisons of interest include pre‐puberty versus puberty, and low IQ versus normal IQ.  

Knowledge about the effectiveness and safety of SSRIs for adult autism would be best served by the conduct of at least one adequately powered RCT of a commonly prescribed drug such as fluoxetine.

Comparison between trials in all age groups would be aided by the use of a core battery of standard outcome measures. As a minimum we recommend a measure of global functioning (for example, CGI); a measure of repetitive and stereotyped behaviours (for example, Repetitive Behavior Scale ‐ Revised); a measure of disruptive behaviour (for example, Aberrant Behavior Checklist), and a measure of obsessive‐compulsive symptoms (for example, Yale‐Brown Obsessive Compulsive Scale).

If short‐term benefit is established in acute trials in the future for one or more key clinical outcomes, then sustained benefit could be explored through the use of a relapse prevention trial conducted over 12 to 18 months. This is relevant as treatments directed to autism tend to be long‐term. A relapse prevention trial also affords the opportunity to obtain systematic adverse event data over a longer period.

Background

Description of the condition

Autism spectrum disorders (ASD) are characterised by qualitative impairment in social interaction and communication skills, as well as stereotypic behaviours and restricted activities and interests. The term ASD has been used commonly in clinical practice, but was only recently introduced into mainstream disease classification systems with the publication of the Diagnostic and Statistical Manual of Mental Disorders fifth edition (DSM‐5) (APA 2013). ASD encompasses conditions previously classified as Infantile Autism (APA 1980), Autistic Disorder (APA 1987; APA 1994) and Classical Autism (WHO 1993). Also included within ASD are conditions previously classified as Pervasive Developmental Disorder ‐ Not Otherwise Specified (PDD‐NOS), 'other pervasive developmental disorders', 'pervasive developmental disorder, unspecified', Asperger syndrome or Asperger disorder and atypical autism.

Estimates of the prevalence of autism using the DSM‐III, DSM‐IIIR, DSM‐IV or International Classification of Diseases (ICD‐10) diagnostic classification systems, from published literature up to April 2004, vary between 1 and 40 per 10,000 and for any ASD between 3 and 82 per 10,000 (Williams 2006). Publications from 2006 have estimated the prevalence of any ASD at between 1.4 and 212 per 10,000 (CDC 2012; Elsabbagh 2012). Males are affected about four times more frequently than females. Problems usually present in early childhood and continue throughout life. Autism places a considerable burden of care on the family and society. Follow‐up studies have found that only 3% to 10% of people with autism are able to live independently as adults (Howlin 2004; Billstedt 2005).

Description of the intervention

Therapies for autism spectrum disorders (ASD)

The heterogeneous nature of problems seen within the autism spectrum means that it is often difficult to be sure which individuals will benefit from the many available therapies. It is also likely that different timing of therapy in relation to age and onset of problems will change outcomes. Many therapies are invasive, time‐consuming and/or expensive and little is known about their potential to cause harm. Pharmacological treatments have been used most commonly as adjuncts to behavioural intervention to target specific symptoms and behaviours. These treatments have been associated with reductions in sleep disturbance, mood disorder, poor attention/concentration and self harm or aggression towards others (Gringras 2000; King 2006).

Selective serotonin reuptake inhibitors (SSRIs)

While antidepressants were the most commonly prescribed medications for ASD, and the fastest growing therapeutic class for this indication through the early 2000s (Aman 2005; Oswald 2007), the trend has changed with recent studies finding neuroleptics (Mandell 2008; Rosenberg 2010) and even stimulants (Rosenberg 2010) being prescribed more often. The prescribing of SSRI drugs to children and adolescents for any indication has been curtailed since the Committee on Safety in Medicines (UK) and the Food and Drug Administration (USA) released safety warnings in 2003 and 2004 respectively, concerning an increased risk of suicide‐related behaviours associated with these medications (Murray 2005; Nemeroff 2007).

How the intervention might work

Increased rates of platelet serotonin transport and levels of whole blood and platelet serotonin (5‐hydroxytryptamine, 5‐HT) have been reported in people with ASD (Cook 1996). Serotonin is linked to the mediation of several psychological processes, many of which are altered in ASD, including mood, social interaction, sleep, obsessive‐compulsive behaviours and aggression (Saxena 1995). It is therefore plausible that inhibition of serotonin reuptake will result in improvement of ASD symptoms.

Why it is important to do this review

Several studies have reported improvements following administration of SSRIs. However, the participant numbers are small, especially for children, and serious side effects including increases in maladaptive behaviours, urinary retention and seizures are reported (Branford 1998). A recent larger study of children showed no improvement from SSRIs (King 2009). To our knowledge, no drug authority has specifically approved the use of SSRIs for autism. The prescribing of SSRIs for autism is therefore either 'off‐label' or is directed to an associated indicated disorder such as obsessive‐compulsive disorder (OCD) or depression.

Regarding indications and prescribing for children, there are between‐country variations. The FDA has approved (allowed the marketing of) sertraline in children six years and older, fluoxetine in children seven years and older, and fluvoxamine in children eight years and older, for the treatment of OCD. The FDA has approved fluoxetine in children eight years and older and escitalopram in adolescents 12 to 17 years for the treatment of depression. In the UK, the Commission on Human Medicines (formerly the Committee on Safety of Medicines) contraindicates all antidepressants other than fluoxetine for the treatment of depression in children and adolescents. In Australia, fluvoxamine has been given a specific indication of OCD in children eight years and over, while prescribers are urged to exercise caution in prescribing other SSRIs for children under the age of 18 years.

A up‐to‐date systematic review of SSRIs is required to assess the evidence of efficacy and harm when used to treat ASD. This review is an update ,of a previously published review (Williams 2010).

Objectives

To determine if treatment with SSRIs:
1. improves the core features of ASD (social interaction, communication and behavioural problems);
2. improves other non‐core aspects of behaviour or function such as self‐injurious behaviour;
3. improves the quality of life of adults or children and their carers;
4. has short‐ and long‐term effects on outcome;
5. causes harm.

Methods

Criteria for considering studies for this review

Types of studies

Trials were eligible for inclusion in the review if the assignment of study participants to intervention or control group was random.

Types of participants

Inclusion was limited to individuals with a diagnosis of an ASD defined using DSM‐IV or ICD‐10 or equivalent as a Pervasive Developmental Disorder, excluding Rett syndrome and Childhood Disintegrative Disorder. Diagnosis must have been made using a standardised diagnostic instrument (Childhood Autism Rating Scale (CARS), Autism Diagnostic Interview‐Revised (ADI‐R), Autism Diagnostic Observation Schedule (ADOS), Diagnostic Interview for Social and Communication Disorders (DISCO)) or by using established diagnostic criteria (ICD‐10, DSM‐IV). No age limits were applied.

Types of interventions

Oral SSRIs, regardless of dosage used or frequency of administration. The control treatment must be a placebo.

Types of outcome measures

Primary outcomes

  1. Core features of ASD, that is, social interaction, communication and behavioural problems including stereotypy or restricted, repetitive patterns of behaviour, interests or activities

Outcome measured by standardised diagnostic assessment instruments (CARS, ADI‐R, ADOS, DISCO) or assessment tools for social communication and repetitive and restricted behaviours.

Secondary outcomes

  1. Non‐core aspects of behaviour and function such as sleep disturbance, self mutilation, aggression, attention and concentration problems, and gastrointestinal function;

  2. Global assessment of health and function;

  3. Quality of life for the individual or their family;

  4. Adverse events.

Outcomes measured by:

  • standardised communication assessments;

  • quality of life questionnaires;

  • rating scales of emotions and behaviour, including depression, anxiety, aggression, obsessive‐compulsive behaviour;

  • global impression rating scales;

  • other health outcome rating scales.

We intended to examine short‐ (up to three months), medium‐ (three to 12 months) and long‐term (greater than 12 months) outcomes if data were available. In future updates, if there are data of sufficient quality to generate a 'Summary of findings' table, we will report the primary outcome (core features of ASD), as well as the secondary outcomes 'global assessment of health and function' and 'adverse events'.

Search methods for identification of studies

This update is based on database searches run in March 2012 and updated in March 2013 and which cover the period since the search date of the previous version of this review (December 2009). We used the most recent version of the Cochrane highly sensitive search strategy for identifying randomised trials in Ovid MEDLINE (Lefebvre 2008). We also adapted the original search strategies for the Education Resource Information Center (ERIC) and Sociological Abstracts because these had previously been searched on different platforms.

Electronic searches

We searched the following databases.

  1. Cochrane Central Register of Controlled Trials (CENTRAL), 2013, Issue 2, part of The Cochrane Library, last searched 16 March 2013

  2. Ovid MEDLINE, 1950 to March Week 1 2013, last searched 16 March 2013

  3. EMBASE (Ovid), 1980 to 2013 Week 11, last searched 16 March 2013

  4. CINAHL (EBSCO), 1937 to current, last searched 16 March 2013

  5. PsycINFO (Ovid), 1806 to March Week 2 2013, last searched 16 March 2013

  6. ERIC, 1966 to current, last searched 16 March 2013

  7. Sociological Abstracts (Proquest), 1952 to current, last searched 16 March 2013

  8. International Clinical Trials Registry Platform (ICTRP), last searched 19 March 2013

  9. ClinicalTrials.gov (CT.gov), last searched 19 March 2013

Detailed search strategies used for this update are in Appendix 1 and the search strategies for the original review are in Appendix 2.

Searching other resources

We also searched bibliographies of articles identified through the search strategy and contacted known experts in the field.

Data collection and analysis

Selection of studies

Two authors screened titles and abstracts from the searches, in 2004, 2006 and 2008 (DW, KW), December 2009 (KW, NS), one author (KW) in 2012 and two authors (AB and MR) in June 2013. We resolved disagreement by consensus and discarded articles that did not fulfil the inclusion criteria. We retrieved potentially relevant articles for full‐text assessment and data extraction.

Data extraction and management

Data were organised using Review Manager 5 software (RevMan 2012). We developed data extraction forms a priori and included information regarding methods, participant details, dose and frequency of SSRI administration, and outcomes. Three pairs of independent review authors [(i) KW and DW, (ii) KW and NS, or (iii) AB and MR] extracted data. No disagreements arose.

Assessment of risk of bias in included studies

Three pairs of authors [(i) KW and DW, (ii) KW and NS, or (iii) AB and MR] independently assessed each included study using the risk of bias criteria outlined in chapter 8 of the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011) without blinding to authorship or source. We compared the assessments for inconsistencies and resolved differences in interpretation by discussion and consensus. We assessed risk of bias according to the following five domains with ratings of low, unclear or high risk of bias.

1. Sequence generation

Was the allocation sequence adequately generated?

Low risk ‐ computer‐generated random numbers, table of random numbers, coin‐tossing or similar.
High risk ‐ day of week, even/odd clinic record number, clinician judgment, participant preference, laboratory test result such as haemoglobin value, or similar.
Unclear risk ‐ insufficient information about the sequence generation process to permit judgment.

2. Allocation concealment

Was allocation adequately concealed?

Low risk ‐ central independent unit, sequentially numbered drug containers or sealed envelopes of identical appearance, or similar.
High risk ‐ alternation or rotation, date of birth, non‐opaque envelopes, open table of random numbers or similar.
Unclear risk ‐ randomisation stated but no information on method used is available.

3. Blinding

Was knowledge of the allocated intervention adequately prevented during the study?

We assessed blinding of participant and personnel and blinding of outcome assessors. We took into account whether the placebo and medication were similar or different and other information provided in the text, such as whether outcome assessors were aware of side effects or participants correctly identified treatment group when available, as well as whether the trialists specifically commented on blinding.

4. Incomplete outcome data

Were incomplete data dealt with adequately by the researchers?

Low risk ‐ no missing outcome data, missing outcome data balanced in numbers across intervention groups and reasons for dropouts and withdrawals described or similar.
High risk ‐ reason for missing outcome data likely to be related to true outcome or similar.
Unclear risk ‐ number or reasons for dropouts and withdrawals not described.

5.      Selective outcome reporting

Are reports of the study free of suggestion of selective outcome reporting?

Low risk ‐ study protocol is available, published reports include all expected outcomes or similar.
High risk ‐ not all of the study's prespecified primary outcomes have been reported, one or more reported primary outcomes were not prespecified or similar.
Unclear risk ‐ insufficient information to permit judgement.

In our coding approach we acknowledge that assessing 'risk of bias' for this issue is very subjective unless trials are available on trial registers. We therefore code as 'unclear' unless trials register information indicates that risk of bias is low or high.

We also explored any other potential sources of bias, such as stopping the study early or extreme baseline imbalance.

Measures of treatment effect

We present effect estimates as provided in individual studies along with one meta‐analysis.

For details of all methods planned in the protocol that were not used in this version of the review but may be used in future updates, see Table 1.

Open in table viewer
Table 1. Methods reported in protocol but not used in this review

Issue

Method

Measures of treatment effect

Continuous data

Where standardised assessment tools generate a score as the outcome measure, we plan to compare the means of these scores and calculate a mean difference for inclusion in meta‐analysis, from data available from trial authors or calculated using methods outlined in Chapters 7 and 9 of the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011). Where studies do not use sufficiently similar instruments to measure an outcome, but the construct measured is similar, we plan to conduct meta‐analysis using standardised mean difference.

Unit of analysis issues

We will assess all included trials to determine the unit of randomisation and whether or not this unit of randomisation was consistent with the unit of analysis. Where cross‐over trials are used, we will extract mean and standard error of paired t‐tests and assess the clinical (pharmacodynamic) suitability of 'wash‐out' period and risk of spillover of drug effect for those who received drug treatment first.

Subgroup analysis

Subgroup analysis will be undertaken if clinically different intervention are identified or there are clinically relevant differences between subject groups:
• age of participants (adult vs paediatric, preschool vs school age)
• diagnostic classification
• medication dose

Sensitivity analysis

Sensitivity analysis will be conducted to assess the impact of study quality on the results of meta‐analyses. For example, we will test to see if studies with high rates of loss to follow‐up or inadequate blinding are more likely to show positive outcomes.

Unit of analysis issues

We did not identify any unit of analysis errors. In all included studies, randomisation, reporting and analysis were per individual participant rather than cluster‐randomised studies.

Dealing with missing data

Where possible, missing data and dropouts were assessed and reported for each included study. Reasons for missing data are provided when reported in trials. Where insufficient data were reported, we contacted the trial authors for further information, but received no replies.

Assessment of heterogeneity

We assessed consistency of results visually and by examining I² (Higgins 2002), a value which describes approximately the proportion of variation in point estimates that is due to heterogeneity rather than sampling error (chance). We will supplement this with a test of homogeneity to determine the strength of evidence that the heterogeneity is genuine.

Assessment of reporting biases

We found insufficient studies to allow for the use of funnel plots to investigate any relationship between effect size and study precision (closely related to sample size).

Data synthesis

Binary data

Since more than two studies presented outcomes from standardised instruments as proportions for adults, the relative risk with 95% confidence intervals, was calculated from meta‐analysis, using a random‐effects model. Number needed to treat was not calculated.

Subgroup analysis and investigation of heterogeneity

Subgroup analysis was not possible because of a lack of data for meta‐analysis.

Sensitivity analysis

Sensitivity analysis was not possible because there were too few data available for meta‐analysis for the same drug therapy, age group and clinical outcome.

Results

Description of studies

Results of the search

The original review, based on searches till December 2009, included seven studies.

The search for this update was run in 2012. From 83 records identified, we found no new published trials. We found one new trial on ClinicalTrials.gov and the author provided methodological information and data when contacted. The most recent search was run in March 2013 from which we identified a further 33 records including one published trial. Thus, we include a total of nine randomised controlled trials (RCTs) with 320 participants in this update.

Of note, we did not search for fenfluramine directly and it is not registered for use as an antidepressant, but was identified as a selective serotonin reuptake inhibitor (SSRI) in electronic searches. We debated its inclusion and chose to retain the two trials that evaluated it to provide a complete clinical picture. The tricyclic antidepressant clomipramine was also identified using this search method and is known to have serotonin uptake inhibitor actions, but was excluded because, unlike fenfluramine, it is included in a systematic review of tricyclic antidepressants and autism spectrum disorders (ASD) (Hurwitz 2012).

Included studies

Nine studies are included in this updated review.

Location

Seven studies were carried out in the USA, one in France and one in Japan. 

Age of participants

Participants were children aged 3 to 17 years in five of the studies (Barthelemy 1989; Leventhal 1993; Sugie 2005; Hollander 2005; King 2009) and adults aged 18 to 60 years in the other four (McDougle 1996; Buchsbaum 2001; NCT00609531; Hollander 2012).

Interventions

The five studies conducted with children used fenfluramine (Barthelemy 1989; Leventhal 1993), fluoxetine (Hollander 2005), fluvoxamine (Sugie 2005) and citalopram (King 2009). The four studies conducted with adults used fluoxetine (Buchsbaum 2001;Hollander 2012), fluvoxamine (McDougle 1996) and citalopram (NCT00609531).  

Diagnostic criteria

Of the five trials in children, two used DSM‐IV‐TR diagnoses of Autistic Disorder, Asperger Disorder or Pervasive Development Disorder ‐ Not Otherwise Specififed (PDD‐NOS) (Hollander 2005; King 2009). One study (King 2009) also required at least moderate severity on the Clinical Global Impression (CGI) illness severity scale and a moderate or greater score for compulsive behaviour items of the Children’s Yale‐Brown Obsessive Compulsive Scales modified for pervasive developmental disorders (CY‐BOCS‐PDD). Both of these studies included children with an intelligence quotient (IQ) in the normal range, with one having an IQ range of 30 to 132 (Hollander 2005) and the other reporting 61% of children with a non‐verbal IQ over 70 (King 2009). One study included children "diagnosed with autism" using DSM‐IV but did not specify how criteria were applied or provide information about intelligence (Sugie 2005). Two studies used DSM‐III criteria for a diagnosis of autism and included children with intellectual impairment with IQ ranges of 16 to 63 (Leventhal 1993) and 30 to 75 (Barthelemy 1989).

Of the four adult studies, one (Buchsbaum 2001) included adults diagnosed using DSM‐IV with autism or Asperger Disorders, and all participants were verbal with an IQ score range of 53 to 119. One study included adults diagnosed with autism using the DSM‐III‐R and ICD‐10 criteria, and individuals included were at least "moderate" in severity using the CGI global severity of illness rating. Both intellectually able and disabled adults were included (McDougle 1996). A third study included adults diagnosed using DSM‐IV‐TR with autistic disorder, a score of 70 or more on the Weschler Abbreviated Scale of Intelligence and high levels of repetitive behaviours with scores of eight or higher at baseline on the CYBOCS‐PDD (NCT00609531). The fourth study (Hollander 2012) included adults diagnosed using DSM‐IV criteria. Individuals were only included if they scored four or higher on CGI Gobal Severity ratings. There was a wide range of intellectual functioning in the participants although the majority (92%) had a full‐scale IQ greater than 70.

Prior treatment rules

One study (Leventhal 1993) included participants who had previously been treated with an SSRI. Trial authors conducted a multi‐centre non‐randomised trial of 30 weeks duration and then extended this trial with a 32 week cross‐over RCT. Thus, there was potential for carry‐over effects from initial treatment, which could result in an underestimate of treatment effect.

Outcome measures

Eighteen different standardised outcome measures were used in the nine included trials (Table 2). Use of a single outcome measure by more than one study was uncommon, but occurred for the clinical global impression (CGI) and the Child's Yale‐Brown Obsessive Compulsive (CY‐BOCS) scales, albeit using different scales and subsections of existing scales. Different versions of assessment tools were used to measure similar outcomes. For example for CGI, the Clinical Global Impression ‐ Improvement scale (CGI‐I) and the Clinical Global Improvement Scale Adapted to Global Autism (CGI‐AD) were used (Guy 1976). Similarly, to measure obsessive‐compulsive behaviour, the Children’s Yale‐Brown Obsessive Compulsive Scales modified for pervasive developmental disorders (CY‐BOCS‐PDD) (Scahill 2006), the Children's Yale‐Brown Obsessive Compulsive scale (CY‐BOCS) (McKay 2003) and the Yale‐Brown Obsessive‐Compulsion Scale (Goodman 1989b; Goodman 1989a) and a modified version of it were used. The different versions of the CGI used measure different constructs, as do the different versions and sections of the obsessive‐compulsion scales used.

Open in table viewer
Table 2. Outcome measures used in included trials

Outcome measure

NCT00609531

Barthelemy

King

Buchsbaum

Hollander

2005

Leventhal

McDougle

Sugie

Hollander 2012

Core features of autism

1

Behavioural Assessment Scale

x

2

Behaviour Summarized Evaluation Scale

x

3

Repetitive Behavior Scale–Revised

x

x

Obsessive‐compulsive behaviour

4

Yale‐Brown Obsessive Compulsive Scale (Y‐BOCS)

x (combined obsession and compulsion score)

x (combined obsession and compulsion score)

x (obsession and compulsion subscales reported separately)

x (compulsion subscale only)

x Modified Y‐BOCS used (combined and separate subscales reported)

x (compulsion subscale only)

5

CGI‐I for obsessive‐compulsive symptoms

x

x

Anxiety

6

Hamilton Rating Scale for Anxiety (HAM‐D)

x

Depression

7

Hamilton Rating Scale for Depression

x

x

Behaviour

8

Ritvo‐Freeman Real Life Rating Scale

x

x

9

Vineland Adaptive Behaviour Scales

x

10

Clinical Global Impression Scale (CGI)

x CGI‐I

x

CGI‐I

x

CGI‐I

x

CGI‐AD

x

CGI‐I

x

Genotype specific only

x

CGI‐I

11

Aberrant Behavior Checklist (ABC)

x

x (irritability subscale only)

12

Connors' Abbreviated Parent and Abbreviated Teacher Questionnaires

x

Aggression

13

Brown Aggression Scale

x

Other standardised outcomes

14

Merrill‐Palmer Scale of Mental Tests

x

15

Wechsler Intelligence Scale for Children

x

16

Alpern‐Boll Developmental Profile

x

Adverse events

17

Fluoxetine Side Effects Checklist

x

18

Suicidality Subscale, Overt Aggression Scale ‐ Modified

x

Clinical Global Improvement Scale Adapted to Global Autism (CGI‐AD)
Clinical Global Impression ‐ Improvement scale (CGI‐I)
Clinical Global Impression ‐ Severity scale (CGI‐S)

Study design
Parallel study design

Four studies used a parallel design (McDougle 1996; King 2009; NCT00609531; Hollander 2012). For these studies there were no statistical or clinical differences at baseline between intervention and placebo groups for important outcome measures, including the CGI‐severity score (King 2009; Hollander 2012), and the obsessive‐compulsive scales used (McDougle 1998; NCT00609531; Hollander 2012).

Cross‐over study design

Four studies used a cross‐over design (Barthelemy 1989; Buchsbaum 2001; Hollander 2005; Sugie 2005). No data were available prior to the second phase. One study of fluoxetine observed a wash‐out period of four weeks (Hollander 2005) and another study of fluvoxamine for two weeks (Sugie 2005). Different SSRIs have different recommended wash‐out periods, ranging from 15 hours (fluvoxamine) to seven to nine days (fluoxetine). Of the four studies that reported using only cross‐over methods, no data was extracted from one because data were only presented for genotypes rather than as treatment‐placebo comparisons (Sugie 2005) and one study used paired t‐test analyses (Buchsbaum 2001) . Two studies reported comparative analyses using ANOVA (Barthelemy 1989; Hollander 2005) and one of these also used mixed regression models for comparative analyses (Barthelemy 1989). Treatment duration ranged from five to 12 weeks (see Characteristics of included studies table). One study (Barthelemy 1989) used a shortened placebo period, where participants received placebo for only one month and active treatment for three months. The trial authors adopted this method because of parent concerns with a lengthy non‐treatment phase. In all studies, follow‐up was short term (12 weeks or less).

Mixed study design

One study had two treatment phases: four weeks placebo, 16 weeks fenfluramine, eight weeks placebo, followed by randomisation of 15 children to a cross‐over phase (Leventhal 1993). This study used paired t‐test analyses and also used ANOVA analyses.

Excluded studies

We excluded seven studies after reading the full paper. Two studies (Gordon 1993; Remington 2001) were trials of clomipramine, a tricyclic antidepressant that has SSRI characteristics but is not classed as such. Four studies were not randomised controlled trials (McDougle 1996; Sanchez 1996; Peral 1999; Doyle 2001). One of these studies (Doyle 2001) was a cost analysis of a treatment not classed as an SSRI. A further study did not involve participants with ASD (Humble 2001).

Risk of bias in included studies

Figure 1 provides a summary of the risk of bias of included studies.


Risk of bias summary: review authors' judgements about each risk of bias item for each included study

Risk of bias summary: review authors' judgements about each risk of bias item for each included study

Allocation

Sequence generation

Sequence generation was adequate in three studies (Sugie 2005; King 2009; Hollander 2012) and unclear in the remaining six.

Allocation concealment

Allocation concealment was adequate in three studies (Sugie 2005; King 2009; NCT00609531). Adequacy of allocation concealment was unclear in six studies (Barthelemy 1989; Leventhal 1993; McDougle 1996; Buchsbaum 2001; Hollander 2005; Hollander 2012). No further unpublished information about allocation concealment was forthcoming from trial authors.

Blinding

Outcome assessors were reported as being blind to treatment allocation in four studies (McDougle 1996; Hollander 2005; King 2009; Hollander 2012) with a fifth study stating that "investigators" were blind (NCT00609531). In the remaining four studies it was not possible to ascertain if outcome assessors were blinded. Of these four studies, two stated that participants and treating physicians were blinded but did not report blinding of outcome assessors (Barthelemy 1989; Buchsbaum 2001) and two studies used the term "double blind" to refer to all blinding (Leventhal 1993; Sugie 2005). Given that the outcome measures used rely on subjective observation and assessment, there is potential for bias where outcome assessors were not adequately blinded to treatment allocation.

Incomplete outcome data

Three studies reported no loss to follow‐up (Barthelemy 1989; McDougle 1996; Buchsbaum 2001). One study (King 2009) reported that 13 of 76 withdrew from the placebo arm and 13 of 73 withdrew from the treatment arm of the trial. Reasons for withdrawal included adverse events (one serious in the treatment group), protocol violation and consent withdrawal. One study (Leventhal 1993) reported one withdrawal prior to the randomised phase and reported no data from this participant. One study (Sugie 2005) excluded one participant due to non‐compliance. One study (Hollander 2012) screened 48 individuals, with 37 being randomly assigned to placebo or treatment group. Thirty‐four participants with postrandomisation data were included in the analysis, including four who dropped out after randomisation but for whom postrandomisation data had been collected. Two of these participants did not comply with the study protocol, one participant discontinued due to relocation and one discontinued due to poor toleration of the treatment. One study (Hollander 2005) reported that, of 62 participants who consented, 18 were excluded for non‐eligibility or non‐compliance and 44 were randomised. Of these, 39 participants were included and their completed outcome data were published, three were excluded due to non‐compliance, one was excluded due to lack of efficacy and one was lost from analysis due to lost records. The exclusion of a participant due to lack of efficacy creates a risk of bias, as does exclusion of those who were not compliant with therapy. One study that had expected to recruit 40 participants only recruited 12, but there was no loss to follow‐up of those recruited (NCT00609531).

Four studies reported not using, or were assessed as not using, an intention‐to‐treat analysis (Leventhal 1993; Hollander 2005; Sugie 2005; Hollander 2012). Four studies reported the use of intention‐to‐treat analysis or did not require any statistical adjustments as they had no losses to follow‐up or changes in treatment allocation (Barthelemy 1989; McDougle 1996; Buchsbaum 2001; King 2009). One study provided raw data prior to publication (NCT00609531).

Selective reporting

The likelihood of selective reporting, that is reporting only those outcomes that showed evidence of treatment effect, was lowest for King 2009, which was registered at the start of the trial. One other trial was registered and provided data for listed outcomes as well as some outcomes not listed at trial registration (NCT00609531). No other trial reported being registered prior to start. Five trials reported negative outcomes (Barthelemy 1989; Leventhal 1993; Buchsbaum 2001; Hollander 2005; King 2009); one reported only positive outcomes (McDougle 1996), and one study reported on all outcome measures assessed (Hollander 2012). One study only reported effectiveness for genetic subgroups (Sugie 2005). All studies reported at least one relevant clinical outcome. We would strongly suspect selective reporting if the primary outcomes to be assessed in this review were not presented in the included trials reports.

To assess selective reporting, we assigned each included study a judgement relating to the risk of bias for that study by answering a prespecified question about the study's adequacy in relation to each of the key criteria. Judgements were 'low risk of bias', 'high risk of bias' and 'unclear risk of bias' (used if the risk of bias was unknown).

Other potential sources of bias

We are unaware of any further potential sources of bias in the included studies.

Effects of interventions

The variability in outcome measures, ages of the study samples and medication precluded meta‐analysis except for one outcome for two trials in adults. Results are presented below by age (children or adults) and compound.

Core ASD features

Children

Three studies used measures that included some of the core features of autism as an outcome (Barthelemy 1989; Sugie 2005; King 2009). Meta‐analysis of core features of autism was not feasible because each study used a different outcome tool.

Citalopram

In the study of citalopram in children, the parent‐rated Repetitive Behavior Scale‐Revised (RBS‐R) (Bodfish 1999) was used, and there were no significant differences in any of the six subscale scores at 12 weeks (P > 0.36 for all) (King 2009).

Fenfluramine

One study (Barthelemy 1989) measured core features of autism using the Behavior Summarized Evaluation scale (BSE). This is a 25‐item scale, of which 11 items deal directly with autistic symptoms. Average scores over four successive one‐week periods were used in analysis. No significant change from baseline or significant difference between treatment and placebo groups was found. No order effect was found in the cross‐over study.

Fluvoxamine

One study (Sugie 2005) measured core features of autism using the Behavior Assessment Scale (BAS), a tool designed by the investigators and provided in the text of the paper. The tool is reported to have a correlation with the Childhood Autism Rating Scale (CARS) (P < 0.0001). However, the primary focus of the paper is the correlation between genetic polymorphisms and response to fluvoxamine. BAS scores were reported for participants based on subgroups assessed by their genotype and it was not possible to determine overall values for treatment and control groups. The authors report that 10 of 18 participants "responded" to treatment.

Adults
Citalopram

One study (NCT00609531) provided results of the RBS‐R, although this was not listed as an outcome on the trials register. There was no significant difference between the citalopram and placebo group for analyses in which the data from an earlier assessment (four weeks) was used for a person in the citalopram group lost to follow‐up for this outcome (mean score citalopram 18.5, placebo 25.2; P = 0.6).

Composite measures of CGI and OCB (child data only)

One study (King 2009) used a composite measure of the CGI improvement scale (CGI‐I) and the CY‐BOCS‐PDD. In this study a CGI‐I score of 1 or 2 and a 25% reduction on the CY‐BOCS‐PDD were required as evidence of improvement. The authors of this study reported that the use of the composite score was a way of "increasing the threshold for positive response". Comparative analysis showed no difference between treatment groups for the composite score at 12 weeks (20.6% for citalopram versus 13.2% for placebo; P = 0.28).

One study (Hollander 2005) used a composite score that included the CGI‐AD and a measure of change of repetitive behaviour, based on the CY‐BOCS. The authors created a composite score by creating "a change score by subtracting the pre‐test CY‐BOCS from the post‐test CY‐BOCS. Negative values on this measure indicate a reduction in repetitive behaviours at post‐test whereas positive scores indicate an increase. This raw change measure was then added to the CGI‐AD measure to augment the overall change in autism severity". Results of the mixed regression analysis indicated a trend towards reduction in this global autism composite improvement measure for participants on fluoxetine as compared to placebo (Z = 1.907, standard error (SE) = 0.703, P = 0.056).

Composite scores used were different and were presented as categorical data in one study and continuous data in the other, so that meta‐analysis was not possible.

Global assessment of health and function

Children
Citalopram

At 12 weeks there was no significant difference in the proportion of CGI‐I scale responders between the citalopram‐treated group (32.9%) and the placebo group (34.2%) (risk ratio (RR) 0.96; 95% confidence interval (CI) 0.61 to 1.51; P = 0.99) (King 2009). Further analysis of this study used the generalised estimating equation method, and found there was no significant difference in the rate of improvement on the CGI‐I scale between the groups (P = 0.94), although both groups improved over time.

Fluoxetine

One study (Hollander 2005) used the CGI‐AD. There was no significant benefit from fluoxetine treatment for this score.

Fluvoxamine

One study (Sugie 2005) of fluvoxamine used the CGI scale to assess improvements in behaviour. However, the results were presented for different genotypes and it was not possible to assess the overall outcome score for treatment and control groups.

Adults
Citalopram

One study (NCT00609531) found no statistical difference between the two groups for CGI‐I at 2, 4, 8 or 12 weeks. For the placebo group the 12‐week (final) CGI score was 2.5 (SE 0.43) and for the citalopram group 3.0 (SE 0.58).

Fluoxetine

Two studies (Buchsbaum 2001; Hollander 2012) used the CGI to measure changes in behaviour. In Buchsbaum 2001 three of six participants showed improvement on the CGI‐I, but as this was a cross‐over trial and first phase data were not reported separately the data could not be included in the meta‐analysis. Continuous outcomes were compared using paired t‐tests and reported for 'baseline' and fluoxetine with no significant change (mean difference ‐1.00; standard deviation (SD) 1.26). However, it is uncertain whether 'baseline' represents the control phase so these data were unsuitable for inclusion in a meta‐analysis. The other study (Hollander 2012) reported a significant difference favouring the treatment group over placebo on the CGI‐I postintervention as rated by treating clinicians, with seven out of 20 responders in the treatment group (35%) and no responders out of 12 in the placebo group (P = 0.03). In the ratings by the independent raters on the CGI‐I improvement was numerically greater in the treatment group (30% versus 0%) compared to placebo group (P = 0.07). The latter data were included in the meta‐analysis, because clinician ratings were prone to unmasking as they were also making dosage changes and gathering side effect information.

Fluvoxamine

McDougle 1996 reported statistically significant improvements in behaviour following treatment with fluvoxamine as assessed using the CGI‐I at 4, 8 and 12 weeks. When presented as a proportion who had shown improvement, 53% of participants in the treatment arm were reported to have improved on the CGI‐I item, while no participants in the placebo arm had improved.

Figure 2 is a forest plot of the comparison of percentage improvement for intervention versus placebo arms. Subgroup meta‐analysis for adults is shown. Only one study of children reported data in this way. As shown, there is a significant improvement for adults as reported by the CGI‐I when data from these two studies are combined.


Forest plot of comparison: Proportion improved for Clinical Global Impression Improvement (CGI‐I)

Forest plot of comparison: Proportion improved for Clinical Global Impression Improvement (CGI‐I)

Non‐core aspects of behaviour and function

Obsessive‐compulsive behaviour

Although stereotypy or restricted, repetitive patterns of behaviour, interests or activities are core features of autism, and may manifest in similar ways to obsessive‐compulsive behaviour, obsessive‐compulsive behaviour per se is not a core feature of autism and is therefore reported here under non‐core features of behaviour.

Children
Citalopram (combined obsession and compulsion score only)

Using CY‐BOCS‐PDD (Scahill 2006), there was no significant difference between the groups in score reduction over time from baseline (mean −2.0 (SD 3.4) points for the citalopram group and −1.9 (SD 2.5) points for the placebo group; P = 0.85) (King 2009). Results for obsessions and compulsions were not reported separately.

Fluoxetine (compulsion score only)

One study (Hollander 2005) used the compulsions questions of the CY‐BOCS as their participants were aged 5 to 16 years, and reported no statistically significant difference between groups (effect size changes were mean of ‐1.3 for phase 1 and ‐0.6 for phase 2).

Although both studies in children (Hollander 2005; King 2009) report no statistically significant change on the CY‐BOCS, presentation of different components of the scales (obsession and compulsion as one score or compulsion score only) meant that available data were not suitable for meta‐analyses.

Adults
Citalopram

This study (NCT00609531) used the CY‐BOCS as the measure and at 12 weeks follow‐up there was no statistically significant difference between the two groups with the mean for the citalopram group of 12.1 and placebo of 10.5 (P = 0.7), with baseline scores for the two groups similar at 14.3 and 14.2 respectively. Only data for overall score, not obsession or compulsion score separately, were available.

Fluoxetine

One study used the full adult version of the tool, reporting a statistically significant improvement in obsessions (P = 0.03) but not compulsions (P = 0.86) and a four‐point difference, favouring treatment groups, that was not statistically significant for the overall score (P = 0.06) (Buchsbaum 2001). As reported for the CGI outcome, uncertainty about whether 'baseline' represents the control phase meant that available data were unsuitable for inclusion in a meta‐analysis. One study (Hollander 2012) used the compulsion sub‐scale of the Yale‐Brown Obssessive‐Compulsive Scale. A significantly greater reduction in scores on the compulsion subscale was found in the treatment group compared to the placebo group (P = 0.005). In the CGI measure of improvement in obsessive‐compulsive symptoms the treatment group had more improved outcomes on overall ratings of repetitive behaviours as rated by clinicians (P = 0.03). In the ratings by the independent evaluators there was no statistically significant difference between the treatment and placebo groups for CGI‐I rating in obsessive‐compulsive symptoms (P = 0.25).

Fluvoxamine

One study of fluvoxamine (McDougle 1996) used a modified version of the Yale‐Brown Obsessive‐Compulsion Scale. There was no significant difference in baseline scores between treatment and control groups. Fluvoxamine was reported to show a treatment benefit compared with placebo (mean difference ‐8.2, 95% CI ‐13.92 to ‐2.48). Sample size was small (N = 30). Statistically significant improvements in both obsession (P < 0.02) and compulsion (P < 0.02) scores were reported at eight weeks and also at 12 weeks (obsession P < 0.02; compulsion P < 0.001).

Both studies (McDougle 1996; Buchsbaum 2001) reported improvement in obsessions, as scored using the Yale‐Brown Obsessive‐Compulsion Scale. One study (McDougle 1996) also reported improvement in compulsions and the combined obsession‐compulsion score.

Behaviour
Citalopram

Of the five subscales of the Aberrant Behavior Checklist–Community version only the irritability scale achieved statistical significance (without any correction for multiple comparisons) from baseline to week 12, and the difference in change scores was small (2.27 points favouring the citalopram group) (King 2009).

Fenfluramine

One study (Leventhal 1993) used the Ritvo‐Freeman Real Life Rating Scale to assess possible improvements in behaviour. The complex arrangement of placebo and treatment phases, including two cross‐overs, made the data from this trial difficult to interpret. To ensure that there was no carry‐over effect or learning of responses from repeat administration of the outcome measures, outcome data from the first phase only were used. Overall, there was no significant improvement in behaviour (mean fenfluramine 0.73, SD 0.11; mean placebo 0.80, SD 0.15). There was a significant improvement reported in motor abnormalities and on parent reports of hyperactivity (P values not reported).

Fluoxetine

One study (Hollander 2012) used the irritability subscale of the Aberrant Behaviour Checklist. No statistically significant difference was found between placebo and treatment groups.

Anxiety (adult data only)
Fluoxetine

One study (Buchsbaum 2001) used the Hamilton Rating Scale for Anxiety and reported significant improvement in the treatment group compared with the control group after eight weeks treatment (mean difference 4.50, SD 3.51, P = 0.03). Sample size was very small (N = 6).

Depression (adult data only)
Fluoxetine

Two studies (Buchsbaum 2001; Hollander 2012) used the Hamilton Rating Scale for Depression (HAM‐D). Buchsbaum 2001 reported no significant benefit in the treatment group compared with the control group (mean difference 3.83, SD 3.87, P = 0.06). In the Hollander 2012 study, participant scores on the HAM‐D were not reported as an outcome measure but were used to monitor severity and patterns of depression in both groups. No participants from either group reported clinical signs of depression.

Aggression (adult data only)
Fluvoxamine

One study (McDougle 1996) reported using the Brown Aggression Scale as an outcome measure. Fluvoxamine was significantly better than placebo at reducing aggression (F = 4.57, P < 0.03).

No other non‐core behaviour outcomes, such as sleep or self mutilation, were reported.

Adverse effects

Children
Citalopram

Significantly more children in the citalopram‐treated group had one or more emergent adverse events compared to placebo (97.3% versus 86.8%, P = 0.03) (King 2009), with adverse events recorded at each bi‐weekly visit using the Safety Monitoring Uniform Report Form, a semi‐structured review of body systems (Greenhill 2004). One child who had not previously suffered seizures experienced a prolonged seizure with loss of consciousness, and required emergency hospitalisation. Although citalopram treatment was suspended, after withdrawal from the trial the child continued to have frequent seizures.

Fenfluramine

One study (Barthelemy 1989) reported that one week after treatment at 1.5 mg/kg the dosage had to be reduced due to adverse effects in four children. There were two cases of increased withdrawal and sadness and two cases of increased stereotypies. Dosage was increased after one month in all but one child with no recurrence of adverse symptoms. Four children experienced poor appetite in the first two weeks of treatment and four children displayed irritability in the second month. Mean weight significantly decreased in the treatment group (P < 0.02) in the first month of treatment, but stabilised by the second month and returned to normal one month post‐treatment.

One study (Leventhal 1993) reported similar weight loss in the first treatment phase, with resolution by the second period of fenfluramine administration. No further assessment of adverse effects was reported.

Fluoxetine

One study (Hollander 2005) used a side effects symptom checklist. There were no significant differences recorded in frequency or severity of adverse effects between children in the treatment or control groups. There was no significant difference between treatment and control groups on the suicide subscale of the Overt Aggression Scale. Six of 37 subjects had their dosage reduced due to agitation, and two of 36 had a "dosage reduction" while on placebo.

Fluvoxamine

One study (Sugie 2005) used only blood biochemistry to evaluate adverse effects. No significant differences were reported between treatment and control groups.

Adults
Citalopram

Four out of six participants who received citalopram reported side effects including apathy (1), sedation (2), decreased sexual interest (1) and flatulence (1) compared to two of six participants in the placebo group who reported dry lips (1) and upper gastrointestinal disturbance associated with over‐eating (1) (NCT00609531).

Fluoxetine

One study (Buchsbaum 2001) did not report assessment of any adverse effects. This small study of six adults was primarily focused on cerebral metabolism. One study (Hollander 2012) assessed frequency of side effects over the 12‐week period and found relatively few side effects. However, a total of 1.4 side effects were reported per participant in the treatment group compared to 0.6 side effects per participant in the placebo group. No significant difference in the HAM‐D suicide items was found between the treatment and placebo groups.

Fluvoxamine

One study (McDougle 1996) of adult participants reported that fluvoxamine was well tolerated. Three participants in the treatment group and one in the control group reported nausea. Two participants in the treatment group and one in the control group reported moderate sedation. All adverse effects were recorded in the first two weeks of treatment. There were no recorded anticholinergic adverse effects and no significant changes in pulse or blood pressure. No electrocardiographic changes were related to fluvoxamine. No seizures or dyskinesias were reported.

Quality of life

No study used any standardised measure of quality of life.

Long term outcomes

No study recorded outcome beyond the length of the trial duration, with the exception of Barthelemy 1989 who monitored weight loss (see adverse effects).

Discussion

Summary of main results

People with ASD are a heterogeneous group. Studies included in this review included children and adults covering a wide age range, diagnosed using different classification systems and assessment procedures, and with different levels of intellectual ability and problem severity. Despite these differences, there is consistency of findings for the studies conducted in children and for those conducted in adults. There is no evidence of benefit for children, based on one large study of citalopram with low risk of bias and from four smaller studies. In adults, only evidence from small studies with unclear risk of bias is available to date, which report significant improvements in clinical global impression (fluvoxamine and fluoxetine), obsessive‐compulsive behaviours (fluvoxamine), anxiety (fluoxetine) and aggression (fluvoxamine).

Treatment with selective serotonin reuptake inhibitors (SSRIs) may cause various adverse effects. One study reported significantly more adverse events in children on citalopram compared to placebo and one serious adverse event, a prolonged seizure (King 2009). Both studies of fenfluramine reported adverse effects in children, including withdrawal and sadness that prompted dosage changes (Barthelemy 1989), and weight loss (Barthelemy 1989; Leventhal 1993). With monitoring, dose adjustment and time, all but one of these adverse effects were resolved. No significant differences were reported for side effects in children in the treatment or placebo group for fluoxetine (Hollander 2005) and little information was available for side effects in children in the fluvoxamine study (Sugie 2005). The adult studies (fluvoxamine and fluoxetine) reported that treatment was generally well tolerated.

Overall completeness and applicability of evidence

To date, only one randomised controlled trial (RCT) each for fluvoxamine and citalopram, and two RCTs for fluoxetine are available for adults. In children, one trial each of fluoxetine, fluvoxamine and citalopram is available and two trials of fenfluramine. Most trials were small and there was considerable variation in type and severity of autism and participant intelligence quotient (IQ) between studies. As such, this evidence is not complete and applicability needs to be assessed for each study individually.

Quality of the evidence

This review again highlights problems with trial methods found in systematic reviews of other treatments for autism spectrum disorders (ASD) (Sinha 2011; Williams 2005; Jesner 2007). Variations in the clinical profile of ASD trial participants, such as the age of participants, their IQ, the severity of their problems and whether they have the problems that the treatment is suggested to ameliorate, are likely to lead to differences in treatment effectiveness. It is not yet known whether these factors influence the effectiveness of a treatment under investigation independently or as inter‐related factors. It is also possible that some measures are suitable for measuring change in participants of some ages and not others, or that they accurately measure an outcome for individuals with one severity of ASD or IQ but not for others. This means that meaningful interpretation of the variations in reported outcomes from the studies included in this review is not straightforward.

This review details the findings of nine RCTs. Two trials each evaluated the effectiveness of fenfluramine and fluvoxamine and citalopram. Three trials have evaluated the effectiveness of fluoxetine. In one multicentre study the sample size was over 100, but the next largest study recruited 39 participants. Small sample sizes increase the likelihood of type II error, that is, that no significant change will be found where one exists. Meta‐analysis can address this where sufficient studies use the same outcome measures, but could only used for one outcome in this review. Exacerbating the above problem of small individual trial sample size and the ability to use meta‐analyses to overcome this, is the use of a variety of outcome measures. Eighteen different outcome measures were used in studies contributing to this review and variations of measures generated for the same outcome (different tool or different items from a given tool) also occurred.

A further concern with outcome measures is their sensitivity to change, and what magnitude of change individuals and families would perceive as sufficient to warrant therapy. Behavioural outcomes such as sleep disturbance, self mutilation, attention and concentration problems, and gastrointestinal function were not assessed by any of the trials, nor was quality of life. Consumer involvement in outcome measure selection is important to both generate data that are meaningful to those who use them and to facilitate practice change if clear evidence of effectiveness (or a lack of effectiveness) is found.

All studies reported outcomes until trial completion, (maximum duration 12 weeks), with the exception of weight loss, which was monitored for longer in one trial (Leventhal 1993). The lack of medium‐ and long‐term follow‐up remains a characteristic problem of trials in ASD.

Other SSRIs are used in clinical practice to treat problems associated with ASD. Our review identified no RCTs of sertraline, paroxetine or escitalopram.

Potential biases in the review process

One included trial was found only because it was registered. It is possible that studies exploring SSRIs and ASD have occurred without being registered and have not been published. We did not identify any other potential biases in the review process.

Agreements and disagreements with other studies or reviews

We found no other reviews on this topic.

Risk of bias summary: review authors' judgements about each risk of bias item for each included study
Figuras y tablas -
Figure 1

Risk of bias summary: review authors' judgements about each risk of bias item for each included study

Forest plot of comparison: Proportion improved for Clinical Global Impression Improvement (CGI‐I)
Figuras y tablas -
Figure 2

Forest plot of comparison: Proportion improved for Clinical Global Impression Improvement (CGI‐I)

Comparison 1 Proportion improved for Clinical Global Impression Improvement (CGI‐I), Outcome 1 Clinical Global Impression ‐ Improvement (CGI‐I).
Figuras y tablas -
Analysis 1.1

Comparison 1 Proportion improved for Clinical Global Impression Improvement (CGI‐I), Outcome 1 Clinical Global Impression ‐ Improvement (CGI‐I).

Table 1. Methods reported in protocol but not used in this review

Issue

Method

Measures of treatment effect

Continuous data

Where standardised assessment tools generate a score as the outcome measure, we plan to compare the means of these scores and calculate a mean difference for inclusion in meta‐analysis, from data available from trial authors or calculated using methods outlined in Chapters 7 and 9 of the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011). Where studies do not use sufficiently similar instruments to measure an outcome, but the construct measured is similar, we plan to conduct meta‐analysis using standardised mean difference.

Unit of analysis issues

We will assess all included trials to determine the unit of randomisation and whether or not this unit of randomisation was consistent with the unit of analysis. Where cross‐over trials are used, we will extract mean and standard error of paired t‐tests and assess the clinical (pharmacodynamic) suitability of 'wash‐out' period and risk of spillover of drug effect for those who received drug treatment first.

Subgroup analysis

Subgroup analysis will be undertaken if clinically different intervention are identified or there are clinically relevant differences between subject groups:
• age of participants (adult vs paediatric, preschool vs school age)
• diagnostic classification
• medication dose

Sensitivity analysis

Sensitivity analysis will be conducted to assess the impact of study quality on the results of meta‐analyses. For example, we will test to see if studies with high rates of loss to follow‐up or inadequate blinding are more likely to show positive outcomes.

Figuras y tablas -
Table 1. Methods reported in protocol but not used in this review
Table 2. Outcome measures used in included trials

Outcome measure

NCT00609531

Barthelemy

King

Buchsbaum

Hollander

2005

Leventhal

McDougle

Sugie

Hollander 2012

Core features of autism

1

Behavioural Assessment Scale

x

2

Behaviour Summarized Evaluation Scale

x

3

Repetitive Behavior Scale–Revised

x

x

Obsessive‐compulsive behaviour

4

Yale‐Brown Obsessive Compulsive Scale (Y‐BOCS)

x (combined obsession and compulsion score)

x (combined obsession and compulsion score)

x (obsession and compulsion subscales reported separately)

x (compulsion subscale only)

x Modified Y‐BOCS used (combined and separate subscales reported)

x (compulsion subscale only)

5

CGI‐I for obsessive‐compulsive symptoms

x

x

Anxiety

6

Hamilton Rating Scale for Anxiety (HAM‐D)

x

Depression

7

Hamilton Rating Scale for Depression

x

x

Behaviour

8

Ritvo‐Freeman Real Life Rating Scale

x

x

9

Vineland Adaptive Behaviour Scales

x

10

Clinical Global Impression Scale (CGI)

x CGI‐I

x

CGI‐I

x

CGI‐I

x

CGI‐AD

x

CGI‐I

x

Genotype specific only

x

CGI‐I

11

Aberrant Behavior Checklist (ABC)

x

x (irritability subscale only)

12

Connors' Abbreviated Parent and Abbreviated Teacher Questionnaires

x

Aggression

13

Brown Aggression Scale

x

Other standardised outcomes

14

Merrill‐Palmer Scale of Mental Tests

x

15

Wechsler Intelligence Scale for Children

x

16

Alpern‐Boll Developmental Profile

x

Adverse events

17

Fluoxetine Side Effects Checklist

x

18

Suicidality Subscale, Overt Aggression Scale ‐ Modified

x

Clinical Global Improvement Scale Adapted to Global Autism (CGI‐AD)
Clinical Global Impression ‐ Improvement scale (CGI‐I)
Clinical Global Impression ‐ Severity scale (CGI‐S)

Figuras y tablas -
Table 2. Outcome measures used in included trials
Comparison 1. Proportion improved for Clinical Global Impression Improvement (CGI‐I)

Outcome or subgroup title

No. of studies

No. of participants

Statistical method

Effect size

1 Clinical Global Impression ‐ Improvement (CGI‐I) Show forest plot

3

Risk Ratio (M‐H, Random, 95% CI)

Subtotals only

1.1 Adults

2

62

Risk Ratio (M‐H, Random, 95% CI)

12.58 [1.77, 89.33]

1.2 Children

1

149

Risk Ratio (M‐H, Random, 95% CI)

0.96 [0.61, 1.51]

Figuras y tablas -
Comparison 1. Proportion improved for Clinical Global Impression Improvement (CGI‐I)