Estimulación magnética transcraneal (EMT) para la esquizofrenia
Resumen
Antecedentes
Los pacientes con esquizofrenia a menudo presentan síntomas que no logran responder completamente a los fármacos antipsicóticos. La estimulación magnética transcraneal (EMT) se ha propuesto como un nuevo tratamiento para los pacientes con esquizofrenia, especialmente los que presentan alucinaciones auditivas persistentes.
Objetivos
Valorar los efectos de la EMT sola, comparada con la EMT simulada o con "tratamiento estándar" y otras intervenciones, para la reducción de los síntomas psicóticos asociados con la esquizofrenia.
Métodos de búsqueda
Se hicieron búsquedas en el registro de ensayos del Grupo Cochrane de Esquizofrenia (Cochrane Schizophrenia Group) (junio 2006, junio 2008, abril 2013). Este registro se ha compilado mediante búsquedas metódicas en MEDLINE, EMBASE, BIOSIS, CINAHL, Dissertation abstracts, LILACS, PSYNDEX, PsycINFO, RUSSMED y Sociofile, y se ha complementado con búsquedas manuales en revistas relevantes y numerosas actas de congresos.
Criterios de selección
Se incluyeron todos los ensayos controlados aleatorios que reclutaron al menos cinco participantes y que compararon EMT con EMT simulada u otro tratamiento para los pacientes con esquizofrenia.
Obtención y análisis de los datos
Los datos se extrajeron de forma independiente. Para los datos dicotómicos se calcularon los riesgos relativos (RR) y sus intervalos de confianza (IC) del 95%. Para los datos continuos, se calcularon las diferencias de medias (DM) y el IC del 95%. Se usó un modelo de efectos fijos. La calidad general de las pruebas se evaluó mediante el enfoque GRADE.
Resultados principales
En la revisión se incluyeron 41 estudios con 1473 participantes. Se encontraron diferencias significativas a favor de la EMT temporoparietal en comparación con la EMT simulada para el estado global medido con la escala CGI (siete ECA, n = 224, DM ‐0,5; IC del 95%: ‐0,76 a ‐0,23; pruebas de muy baja calidad) y los síntomas positivos medidos con la escala PANSS (cinco ECA, n = 127, DM ‐6,09; IC del 95%: ‐10,95 a ‐1,22; pruebas de muy baja calidad). Los participantes presentaron significativamente más cefaleas en el grupo de EMT temporoparietal (diez ECA, n = 392, RR 2,65; IC del 95%: 1,56 a 4,50; pruebas de muy baja calidad). Sin embargo, no más participantes del grupo EMT abandonaron temprano el estudio en comparación con los de la EMT simulada (pruebas de muy baja calidad). El estado cognitivo se evaluó mediante 39 medidas diferentes, y todas fueron ambiguas (pruebas de muy baja calidad).
Solamente se incluyeron dos ensayos que compararon la EMT temporoparietal con tratamiento estándar. En estos ensayos los participantes recibieron antipsicóticos de primera y segunda generación en ambos grupos de tratamiento, por lo que la EMT se utilizó como tratamiento coadyuvante a la medicación. No se encontraron diferencias significativas en el número de participantes que mostraron mejoría clínica en el estado global (un ECA, n = 100, RR 1,19; IC del 95%: 0,91 a 1,57) o que abandonaron temprano el estudio (dos ECA, n = 140, RR 0,33; IC del 95%: 0,08 a 1,46) (para ambos pruebas de muy baja calidad). Ningún estudio informó la puntuación del estado global, el estado mental, el estado cognitivo y los efectos adversos.
Para la EMT prefrontal comparada con la EMT simulada, el estado global se midió con tres escalas diferentes, y todas presentaron resultados dudosos (pruebas de muy baja calidad). No fue posible agrupar los datos del estado mental con la escala PANSS debido a la gran heterogeneidad. El estado cognitivo se evaluó mediante 19 medidas diferentes, y 15/19 fueron ambiguas (pruebas de muy baja calidad ). La EMT prefrontal causó más cefaleas (seis ECA, n = 164, RR 2,77; IC del 95%: 1,22 a 6,26; pruebas de muy baja calidad), pero no hubo diferencias en el número de participantes que abandonaron temprano el estudio (pruebas de muy baja calidad). Ningún estudio informó datos de la mejoría clínica.
Se encontró una diferencia significativa a favor de la EMT con estimulación theta en salvas prefrontal en comparación con la EMT simulada para el estado mental en la escala PANNS (tres ECA, n = 108, DM ‐5,71; IC del 95%: ‐9,32 a ‐2,10; pruebas de muy baja calidad). No se encontraron diferencias para la mejoría clínica, el estado cognitivo, el número de cefaleas y el abandono temprano del estudio (pruebas de muy baja calidad).
Ninguno de los estudios incluidos informó satisfacción con la atención.
Conclusiones de los autores
Según esta revisión, no hay pruebas suficientes para apoyar o refutar el uso de EMT para tratar los síntomas de la esquizofrenia. Aunque algunas pruebas indican que la EMT, y en particular la EMT temporoparietal, puede mejorar ciertos síntomas (como las alucinaciones auditivas y los síntomas positivos de esquizofrenia) en comparación con la EMT simulada, los resultados no fueron suficientemente consistentes para no ser ambiguos a través de las medidas de evaluación utilizadas. No hubo pruebas suficientes para indicar cualquier efecto beneficioso agregado con la EMT utilizada como tratamiento coadyuvante a los fármacos antipsicóticos.
La calidad general de las pruebas se calificó como muy baja debido al riesgo de sesgo, lo que se acompañó de imprecisión en las estimaciones debido al número relativamente pequeño de participantes en los estudios. Por lo tanto, se debe considerar mejorar la calidad de los procesos de los ensayos, así como la calidad del informe de los ensayos de EMT en curso y futuros, para facilitar que se hagan valoraciones futuras exactas en la evaluación del riesgo de sesgo. Las diferencias en las técnicas de EMT con respecto a la intensidad de la estimulación, la duración de la estimulación, las áreas cerebrales estimuladas y las variaciones en el diseño de la EMT simulada contribuyeron a la heterogeneidad de los resultados de los estudios y limitaron su interpretación y aplicabilidad. Además, los ensayos evaluaron sus resultados con diversas escalas y los datos utilizables fueron limitados. Por lo tanto, para evaluar mejor los efectos del tratamiento con EMT en los pacientes con esquizofrenia, se recomienda el uso de protocolos de tratamiento y medidas de resultado estandarizados.
PICO
Resumen en términos sencillos
Estimulación magnética transcraneal (EMT) para el tratamiento de la esquizofrenia
Pregunta de la revisión
¿La estimulación magnética transcraneal (EMT) es útil para tratar a los pacientes con esquizofrenia?
Antecedentes
La estimulación magnética transcraneal es un tratamiento relativamente nuevo y complejo basado en un dispositivo. La EMT incluye la aplicación cuidadosa de un campo magnético fuerte cerca de la superficie del cuero cabelludo. El dispositivo de EMT administra pulsos magnéticos fuertes y muy breves que estimulan al cerebro y su red de neuronas. La EMT es una técnica no invasiva relativamente indolora que estimula partes del cerebro (la corteza cerebral). Se ha mostrado que la actividad cerebral difiere en los pacientes con esquizofrenia comparados con otras personas.
Los pacientes con esquizofrenia a menudo presentan síntomas como escuchar voces o ver cosas (alucinaciones) que no logran responder completamente a la medicación. La EMT se ha propuesto como un tratamiento nuevo para los pacientes con esquizofrenia, especialmente los que presentan alucinaciones auditivas persistentes. A menudo los fármacos antipsicóticos también tienen efectos secundarios debilitantes como aumento de peso, apatía o falta de estímulo, y temblores. La EMT podría ser un tratamiento alternativo para los pacientes que no lidian adecuadamente con la medicación estándar.
Descripción de los estudios
Se realizó una búsqueda de ensayos en 2013 y 41 estudios controlados aleatorios están ahora incluidos en esta revisión. Los estudios incluyeron pacientes con diagnóstico de esquizofrenia y asignaron al azar a los participantes a recibir EMT temporoparietal, EMT prefrontal, EMT simulada o atención estándar.
Resultados
En este momento no hay pruebas sólidas para apoyar el uso de EMT para tratar la esquizofrenia. Algunas pruebas de muy baja calidad parecen indicar tentativamente que la EMT puede mejorar el estado global y ciertos síntomas como oír voces, en comparación con la EMT simulada. Sin embargo, actualmente los estudios de investigación no son sólidos ni consistentes, ni están lo suficientemente estandarizados, para apoyar cualquier conclusión firme acerca del uso de la EMT para la esquizofrenia.
No hubo pruebas que indicaran que la EMT puede mejorar los síntomas de la esquizofrenia cuando se utiliza junto con el tratamiento estándar de los fármacos antipsicóticos. También hubo limitaciones relacionadas con las diversas técnicas de EMT. Fue difícil comparar los resultados de los estudios en esta revisión porque se utilizaron diversos procedimientos de EMT, diferentes medidas de los síntomas de esquizofrenia, y los datos fueron limitados. Por lo tanto, se requieren estudios de investigación más sólidos y consistentes. Los revisores indican que en el futuro, con más estudios de investigación, es posible que la EMT sea útil para tratar algunos de los síntomas de la esquizofrenia.
Conclusiones de los autores
Summary of findings
| TEMPOROPARIETAL TMS compared to SHAM TMS for schizophrenia | ||||||
| Patient or population: people with schizophrenia | ||||||
| Outcomes | Illustrative comparative risks* (95% CI) | Relative effect | No of Participants | Quality of the evidence | Comments | |
| Assumed risk | Corresponding risk | |||||
| SHAM TMS | TEMPOROPARIETAL TMS | |||||
| Clinical improvement in global state | 0 per 1000 | 0 per 1000 | RR 7 | 46 | ⊕⊝⊝⊝ | |
| Global state score | The mean scores for the sham TMS group ranged from 2.4 to 5.1 | The mean global state score in the intervention groups was | 224 | ⊕⊕⊝⊝ | ||
| Mental state | The mean mental state scores in the sham TMS group ranged from 63.92 to 85.8 | The mean mental state in the intervention groups was | 127 | ⊕⊕⊝⊝ | ||
| Cognitive state | See comment | See comment | Not estimable | 82 | ⊕⊕⊝⊝ | Cognitive state was reported in 3 studies using 39 different measures. Results were equivocal for all measures |
| Adverse effects: general or specific | See comment | See comment | Not estimable | 442 | ⊕⊕⊝⊝ | There were more headaches and jaw and facial contraction in the TMS group. Results for other adverse events ‐ concentration problems, earache, lightheadedness, mild amnesia, restless legs, somatic discomfort, tingling sensation in the arm, worsening hallucinations ‐ were equivocal |
| Adverse effects: Leaving the study early | 152 per 1000 | 118 per 1000 | RR 0.78 | 320 | ⊕⊝⊝⊝ | |
| Satisfaction with care ‐ not reported | See comment | See comment | Not estimable | ‐ | See comment | No studies reported on this outcome |
| *The basis for the assumed risk is the median control group risk across studies for pooled data and the control group risk for single studies. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). | ||||||
| GRADE Working Group grades of evidence | ||||||
| 1Risk of bias: serious ‐ this study had an unclear risk of bias for randomisation, allocation concealment, blinding and incomplete outcome data. Downgraded one level. | ||||||
| TEMPOROPARIETAL TMS compared to STANDARD TREATMENT for schizophrenia | ||||||
| Patient or population: people with schizophrenia | ||||||
| Outcomes | Illustrative comparative risks* (95% CI) | Relative effect | No of Participants | Quality of the evidence | Comments | |
| Assumed risk | Corresponding risk | |||||
| STANDARD TREATMENT | TEMPOROPARIETAL TMS | |||||
| Clinical improvement in global state | 620 per 1000 | 738 per 1000 | RR 1.19 | 100 | ⊕⊕⊝⊝ | |
| Global state score ‐ not reported | See comment | See comment | Not estimable | ‐ | See comment | No studies reported data for this outcome |
| Mental state ‐ not reported | See comment | See comment | Not estimable | ‐ | See comment | No studies reported data for this outcome |
| Cognitive state ‐ not reported | See comment | See comment | Not estimable | ‐ | See comment | No studies reported data for this outcome |
| Adverse effects: general or specific ‐ not reported | See comment | See comment | Not estimable | ‐ | See comment | No studies reported data for this outcome |
| Adverse effects: Leaving the study early | 86 per 1000 | 28 per 1000 | RR 0.33 | 140 | ⊕⊝⊝⊝ | |
| Satisfaction with care ‐ not reported | See comment | See comment | Not estimable | ‐ | See comment | No studies reported data for this outcome |
| *The basis for the assumed risk is the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). | ||||||
| GRADE Working Group grades of evidence | ||||||
| 1Participants received first and second generation antipsychotics in both treatment groups | ||||||
| PREFRONTAL TMS compared to SHAM TMS for schizophrenia | ||||||
| Patient or population: people with schizophrenia | ||||||
| Outcomes | Illustrative comparative risks* (95% CI) | Relative effect | No of Participants | Quality of the evidence | Comments | |
| Assumed risk | Corresponding risk | |||||
| SHAM TMS | PREFRONTAL TMS | |||||
| Clinical improvement in global state ‐ not reported | See comment | See comment | Not estimable | ‐ | See comment | No studies reported data for this outcome |
| Global state score | See comment | See comment | Not estimable | 85 | See comment | 3 small studies measured global state on the CGI, CGI‐S, GAF and SCL‐90, none of which showed a significant treatment effect |
| Mental state | See comment | See comment | Not estimable | 188 | ⊕⊝⊝⊝ | There was very high heterogeneity for this outcome, so we did not pool the data |
| Cognitive state | See comment | See comment | Not estimable | 138 | ⊕⊕⊝⊝ | Cognitive state was reported in 4 studies using 19 different measures. Results were equivocal for all measures apart from 4 |
| Adverse effects: general or specific | See comment | See comment | Not estimable | 199 | ⊕⊕⊝⊝ | There were more headaches and TMS‐related site discomfort or pain in the TMS group. Results for other adverse effects ‐ cognitive difficulties, facial twitching, worsening of pre‐existing akathisia and OCD ‐ were equivocal |
| Adverse effects: Leaving the study early | 106 per 1000 | 126 per 1000 | RR 1.19 | 174 | ⊕⊕⊝⊝ | |
| Satisfaction with care ‐ not reported | See comment | See comment | Not estimable | ‐ | See comment | No studies reported data for this outcome |
| *The basis for the assumed risk (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). | ||||||
| GRADE Working Group grades of evidence | ||||||
| 1Risk of bias: serious ‐ two studies had an unclear risk of bias for randomisation and five for allocation concealment. Five studies had an unclear risk for blinding of participants and two for blinding of outcome assessors and incomplete outcome data. | ||||||
| PREFRONTAL TBS TMS compared to SHAM TMS for schizophrenia | ||||||
| Patient or population: people with schizophrenia | ||||||
| Outcomes | Illustrative comparative risks* (95% CI) | Relative effect | No of Participants | Quality of the evidence | Comments | |
| Assumed risk | Corresponding risk | |||||
| SHAM TMS | PREFRONTAL TBS TMS | |||||
| Clinical improvement in global state | 0 per 1000 | 0 per 1000 | RR 4.06 | 27 | ⊕⊝⊝⊝ | |
| Global state score ‐ not reported | See comment | See comment | Not estimable | ‐ | See comment | No studies reported data for this outcome |
| Mental state | The mean scores for the sham TMS group ranged from 67.6 to 68.4 | The mean mental state in the intervention groups was | 108 | ⊕⊕⊝⊝ | ||
| Cognitive state | See comment | See comment | Not estimable | 39 | ⊕⊕⊝⊝ | This was measured on 2 scale, both showed equivocal results |
| Adverse effects: general or specific | See comment | See comment | Not estimable | 27 | ⊕⊕⊝⊝ | No differences in headaches and sleep disorders |
| Adverse effects: Leaving the study early | 139 per 1000 | 50 per 1000 | RR 0.36 | 76 | ⊕⊕⊝⊝ | |
| Satisfaction with care ‐ not reported | See comment | See comment | Not estimable | ‐ | See comment | No studies reported data for this outcome |
| *The basis for the assumed risk (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). | ||||||
| GRADE Working Group grades of evidence | ||||||
| 1Risk of bias: serious ‐ there was an unclear risk of bias for allocation concealment and blinding | ||||||
Antecedentes
Descripción de la afección
Habitualmente los pacientes con esquizofrenia presentan alucinaciones auditivas (escuchan voces) o delirios (falsas creencias) durante los episodios agudos. Aunque hay varios tratamientos eficaces disponibles, muchos pacientes presentan síntomas intratables que no se recuperan entre los episodios agudos. Además, también se pueden perjudicar la motivación y el comportamiento social (síntomas negativos). Un número relativamente alto de pacientes con esquizofrenia presentan síntomas persistentes a pesar de la farmacoterapia aparentemente suficiente. En algunos casos el fracaso del tratamiento se asocia con la falta de cumplimiento, aunque se entiende que muchos pacientes presentan síntomas persistentes a pesar del tratamiento adecuado. La estimulación magnética transcraneal (EMT) podría ser un tratamiento alternativo en los pacientes que no lidian adecuadamente con el tratamiento estándar.
Descripción de la intervención
La estimulación magnética transcraneal es un tratamiento complejo relativamente nuevo basado en un dispositivo que incluye la aplicación cuidadosa de un campo magnético fuerte cerca de la superficie del cuero cabelludo. El procedimiento es una técnica no invasiva y relativamente indolora para estimular la corteza cerebral y alterar la función neuronal (Chouinard 2003). El dispositivo emplea espirales de alambre aislados específicamente diseñados que administran pulsos magnéticos fuertes y muy breves, que pasan sin dificultad de puntos de referencia superficiales elegidos cuidadosamente hacia las regiones cerebrales subyacentes. El campo magnético induce luego corrientes eléctricas transitorias pequeñas en el circuito neural de los pacientes tratados. Al variar la intensidad, la duración y la frecuencia del campo magnético, los sistemas neuronales se pueden excitar o inhibir mientras persistan los pulsos de corriente en la espiral(Barker 2002).
De qué manera podría funcionar la intervención
Se ha mostrado que la actividad cerebral difiere en los pacientes con esquizofrenia comparados con la actividad cerebral de los pacientes que no tienen esta afección. Mientras la actividad en la corteza temporoparietal (CTP) parece aumentar en los pacientes con esquizofrenia que presentan alucinaciones auditivas (Shergill 2000), la actividad en la corteza prefrontal dorsolateral (CPFDL) parece reducirse en estos pacientes (Weinberger 1996). La reducción de la actividad también parece correlacionarse con síntomas negativos (p.ej. disminución de la motivación y la función social) experimentados por el paciente. Es posible que al normalizar la actividad en estas regiones cerebrales, también mejoren las alucinaciones auditivas y los síntomas negativos.
La estimulación magnética transcraneal se ha aplicado en varios ensayos en dos paradigmas principales: EMT de alta frecuencia y EMT de baja frecuencia. La EMT de baja frecuencia (1 Hz) se aplica habitualmente a la CTP izquierda de los pacientes, con lo que se intenta reducir la actividad cerebral y las alucinaciones auditivas. La EMT de alta frecuencia se aplica a la CPFDL izquierda en un intento por aumentar la actividad y reducir los síntomas negativos. Se considera que la EMT de baja frecuencia inhibe la actividad cortical (Chen 1997) y la EMT de alta frecuencia aumenta generalmente la actividad cortical en las áreas estimuladas (Pascual‐Leone 1998). Se han publicado estudios controlados positivos (Hoffman 2005) y negativos (McIntosh 2004) que utilizan ambos abordajes terapéuticos, y no está claro si la EMT representa un adelanto terapéutico significativo.
En la esquizofrenia hay pruebas de reducción y aumento de la actividad cortical en comparación con los controles no afectados, y en algunos casos la actividad alterada se correlaciona con la presencia de un síntoma conocido de déficit cognitivo. Los estudios han demostrado una asociación entre la actividad del lóbulo temporal y alucinaciones auditivas en los pacientes con esquizofrenia (D'Alfonso 2002; Hoffman 2000; Lee 2005; Poulet 2005). Se ha encontrado que la estimulación activa reduce significativamente las alucinaciones en comparación con la estimulación simulada (Hoffman 2000). No todos los intentos de replicación han apoyado de manera indudable los resultados de Hoffman (McIntosh 2004; Saba 2006b).
Por qué es importante realizar esta revisión
Los brazos placebo de los ensayos de EMT a menudo utilizan tratamientos simulados. Existen limitaciones en este enfoque; no se ha establecido una condición placebo satisfactoria y los individuos pueden no tener expectativas idénticas con la EMT real o simulada. La EMT placebo o simulada debe provocar una sensación en el cuero cabelludo y un ruido idénticos a los de la EMT activa, sin estimulación cortical. Aunque es posible imitar la sensación de ruido, la generación de la sensación en el cuero cabelludo también puede producir una estimulación cortical terapéutica. Al evitar los factores de confusión de la estimulación cortical con la EMT simulada se produce un brazo control del ensayo que habitualmente controla la sensación de ruido pero no la sensación en el cuero cabelludo. En los ensayos que implementan la EMT simulada, no se garantiza el doble cegamiento del observador y el participante y las tasas de eficacia calculadas de la EMT posiblemente estarán afectadas por factores de confusión si los participantes están al tanto de en qué brazo de tratamiento están.
Con las limitaciones de la metodología de los ensayos en mente y a falta de una condición simulada completamente inactiva que imite la EMT real, esta revisión sistemática tiene como objetivo evaluar la base de pruebas actuales de la EMT para el tratamiento de la esquizofrenia. Se deseaba evaluar la eficacia y la seguridad de la EMT, explorar las fuentes de heterogeneidad que podrían explicar los efectos contradictorios positivos y negativos, investigar si se pueden obtener tamaños del efecto agrupados y si son estadísticamente sólidos, y finalmente, proporcionar de ser posible recomendaciones para los estudios de investigación futuros.
Objetivos
Valorar los efectos de la EMT sola comparada con la EMT simulada o con "tratamiento estándar" y cualquier otra intervención de comparación en la reducción de los síntomas psicóticos asociados con la esquizofrenia.
Métodos
Criterios de inclusión de estudios para esta revisión
Tipos de estudios
Todos los ensayos controlados aleatorios relevantes con tamaños de los grupos de al menos cinco. Los ensayos se incluyeron cuando se describieron como "doble ciego" pero se dio a entender que el estudio era aleatorio y los detalles demográficos de cada grupo eran similares. Se excluyeron los estudios cuasialeatorios, como los que se asignaron por días alternados de la semana.
Tipos de participantes
Pacientes con esquizofrenia y psicosis afectivas relacionadas, diagnosticadas según criterios operacionales estandarizados, independientemente de la edad y el sexo.
Tipos de intervenciones
1. Estimulación magnética transcraneal: a cualquier voltaje, frecuencia o carga de estímulo, administrada en la cabeza en cualquiera ubicación
2. EMT simulada: EMT administrada mediante instrumentos falsos o con el espiral aplicado en un ángulo oblicuo, mayor o igual a 45 grados, al cráneo
3. Tratamiento estándar: cualquier tratamiento (incluidos los fármacos antipsicóticos) proporcionado como parte de la atención habitual, definida de cualquier manera
4. Otros tratamientos farmacológicos o no farmacológicos administrados como parte de una intervención experimental. Los ejemplos podrían incluir tratamiento electroconvulsivo (TEC) y terapia cognitivo‐conductual (TCC).
Tipos de medida de resultado
Se clasificaron los resultados en las ocho categorías que se detallan a continuación:
Resultados primarios
1. Estado global
1.1 Mejoría clínica en el estado global (como la definieron los estudios individuales)
1.2 Media de la puntuación de la variable principal de resultado del estado general
1.3 Media del cambio en las puntuaciones del estado global
Resultados secundarios
2. Estado mental
2.1 Mejoría clínica en el estado mental general (como la definieron los estudios individuales)
2.2 Media de la puntuación de la variable principal de resultado del estado mental general
2.3 Media del cambio en las puntuaciones del estado mental general
2.4 Ningún cambio clínicamente importante en los síntomas específicos
2.5 Media de la puntuación de la variable principal de resultado de los síntomas específicos
2.6 Media del cambio en las puntuaciones de síntomas específicos
3. Estado cognitivo
3.1 Mejoría clínica en el estado cognitivo (como la definieron los estudios individuales)
3.2 Media de la puntuación de la variable principal de resultado del estado cognitivo
3.3 Media del cambio en las puntuaciones del estado cognitivo
3.4 Media de la puntuación de la variable principal de resultado específica del estado cognitivo
3.5 Media del cambio en las puntuaciones del estado cognitivo específico
4. Efectos adversos
4.1 Incidencia de efectos adversos, generales o específicos
4.2 Abandono temprano del estudio
4.3 Aceptación medida del tratamiento
4.4 Uso de tratamiento antiparkinsoniano
4.5 Muerte súbita e inesperada
5. Resultados del hospital y los servicios
5.1 Hospitalización de pacientes en la comunidad
5.2 Duración de la estancia hospitalaria
5.3 Gravedad de los síntomas en el momento del alta hospitalaria
5.4 Cambios en el estado de hospitalización (por ejemplo, cambios de la atención informal a la detención formal para atención, cambios en el nivel de observación por el personal de la sala y uso de un ámbito de enfermería aislado)
5.5 Cambios en los servicios prestados por equipos en la comunidad
6. Satisfacción con la atención
6.1 Receptor de la atención
6.2 Cuidadores informales
6.3 Cuidadores profesionales.
7. Resultados económicos
8. Calidad de vida
8.1 Mejoría clínica en la calidad de vida (como la definieron los estudios individuales)
8.2 Media de la puntuación de la variable principal de resultado de la calidad de vida
8.3 Media del cambio en las puntuaciones de la calidad de vida
Tabla "Resumen de los hallazgos"
Se utilizó el enfoque GRADE para interpretar los resultados (Schünemann 2008) y GRADE profiler para importar los datos de la revisión de RevMan 5 (RevMan) para crear las tablas "Resumen de los hallazgos". Estas tablas proporcionan información específica de los resultados en cuanto a la calidad general de las pruebas de cada estudio incluido en la comparación, la magnitud del efecto de las intervenciones examinadas y la suma de los datos disponibles sobre todos los resultados que se consideraron importantes para la atención del paciente y la toma de decisiones. Se incluyeron los siguientes resultados a corto o medio plazo en las tablas "Resumen de los hallazgos":
1. Estado global
1.1 Mejoría clínica en el estado global
1.2 Media de la puntuación de la variable principal de resultado
2. Estado mental
2.1 Media de la puntuación de la variable principal de resultado
3. Estado cognitivo
3.1 Media de la puntuación de cambio
4. Efectos adversos
4.1 Incidencia de efectos adversos, generales o específicos
4.2 Abandono temprano del estudio
5. Satisfacción con la atención
5.1 Receptor de la atención
Results
Description of studies
For a full description of studies please see: Characteristics of included studies and Characteristics of excluded studies.
Results of the search
The search strategy identified 99 reports that were potentially relevant. Agreement about which reports may have been randomised was 100%. In total, we included 41 studies in the review and in the analysis (see Figure 1). One study (Jin 2012) met the inclusion criteria but did not report data in a usable way, and is in the excluded studies table.
Study flow diagram.
Included studies
1. Length of studies
The duration of trials ranged from four days (McIntosh 2004) to 10 weeks (Wing 2012); 26 trials were short (from five days to two weeks), 12 were medium length (three to six weeks) and four trials were long (eight weeks or longer).
2. Design
All but two included studies presented a parallel longitudinal design and two studies had a cross‐over design (McIntosh 2004; Poulet 2005).
3. Participants
Most studies included participants with schizophrenia or schizoaffective disorder according to the DSM‐IV. Of those that did not use DSM‐IV, seven studies diagnosed schizophrenia according to the CCMD‐3 (Gao 2009a; Gao 2009b; Gao 2010; Hao 2008; Liu 2008; Xu 2011; Zheng 2012). Rosenberg 2012 diagnosed according to DSM‐IV‐TR. Bagati 2009, Guse 2013, and Prikryl 2007 used ICD‐10, and De Jesus 2011 OPCRIT 4.0.
In total, 1473 participants are included in the review, and the number of people included in individual studies ranged from 10 (Poulet 2005) to 100 (Liu 2011).
4. Settings
Eleven studies included inpatients (Chen 2011; Cordes 2010; Gao 2009a; Hao 2008; Holi 2004; Klein 1999; Liu 2008; Poulet 2005; Prikryl 2007; Saba 2006a; Zheng 2012), five studies included outpatients (Fitzgerald 2005; Fitzgerald 2008; Mogg 2005; Schneider 2008; Wing 2012) and five studies included both inpatients and outpatients (Bagati 2009; Guse 2013; Novak 2006; Rosenberg 2012; Vercammen 2009a). In 22 studies the setting was either unclear or not reported.
Fourteen studies were carried out in China (Chen 2011; Gao 2009a; Gao 2009b; Gao 2009c; Gao 2010; Hao 2008; Liu 2008; Liu 2011; Ren 2010; Ren 2011; Xu 2011; Yu 2010; Zhang 2010; Zheng 2012), 13 in Europe, including three in the Czech Republic (Klirova 2010; Novak 2006; Prikryl 2007), three in France (Brunelin 2006; Poulet 2005; Saba 2006a), two in Germany (Cordes 2010; Guse 2013), two in the Netherlands (Slotema 2011; Vercammen 2009a), two in the UK (McIntosh 2004; Mogg 2005), and one in Finland (Holi 2004). Of the remainder, six were conducted in the USA (Hoffman 2005; NCT00308997; Schneider 2008) or Canada (Barr 2013; Blumberger 2012; Wing 2012), two in Australia (Fitzgerald 2005; Fitzgerald 2008), two in Brazil (De Jesus 2011; Rosa 2007), two in Israel (Klein 1999; Rosenberg 2012), one in India (Bagati 2009), and one in Korea (Lee 2005).
5. Interventions
5.1 Temporoparietal TMS
Twenty‐two studies used temporoparietal TMS, most using the left temporoparietal region, Lee 2005 also using right temporoparietal TMS, and NCT00308997 using Wernicke's area and right homologous area. Most studies used low‐frequency TMS with 1 Hz at 80 to 110% motor threshold; Hao 2008 and Liu 2008 both used 10 Hz at 110% motor threshold, Saba 2006a used 1 Hz at 20% of motor threshold and Klirova 2010 used 0.9 Hz at 100% motor threshold. Blumberger 2012 included two TMS groups, one with priming TMS of 6 Hz at 90% motor threshold.
In regards to length of TMS stimulations, a wide variety were reported across studies, ranging from five sessions of one minute, with one minute gaps (Saba 2006a) to 12 sessions of 20 minutes each a day (Vercammen 2009a). Blumberger 2012 used MRI‐targeted TMS, and Klirova 2010 and Slotema 2011 both included an MRI‐targeted TMS arm and a non‐targeted TMS arm. De Jesus 2011, Hoffman 2005, and McIntosh 2004 reported using the 10 ‐ 20 EEG electrode position system. Rosenberg 2012 used deep H1 coil TMS with single pulse stimulation, which allows stimulation of deeper brain areas.
5.2 Prefrontal TMS
Nineteen studies used prefrontal TMS, with most using left prefrontal TMS or left dorsolateral prefrontal cortex TMS. Klein 1999 reported using right prefrontal TMS and Barr 2013; Fitzgerald 2008; Ren 2010; Ren 2011 and Wing 2012 reported using bilateral prefrontal TMS. Various stimulations of TMS were administered. Seven studies reported using 10 Hz at 90 to 110% motor threshold, Gao 2009c used 15 Hz at 90% motor threshold, three studies used 20 Hz at 90% motor threshold (Barr 2013; Novak 2006; Wing 2012). Klein used low‐frequency TMS with 1 Hz at 10% above threshold. Two studies (Chen 2011; Zhang 2010) used theta burst stimulation (TBS) TMS, in which 50 Hz are applied in bursts, and Zheng 2012 used three arms of TMS 10 Hz, 20 Hz and TBS (50 Hz).
In terms of stimulation length for each session of TMS, there was much variation reported across the studies. TMS ranged from two trains of one minute with a three‐minute gap (Klein 1999), to 40 trains of 2.5 seconds with a 30‐second gap (Novak 2006). Barr 2013 used MRI‐targeted TMS and Guse 2013 and Poulet 2005 reported using the 10 ‐ 20 EEG electrode position system.
5.3 Sham TMS
For the sham TMS condition a variety of techniques were used. Seventeen studies described using the same stimulation as for active TMS but with the edge resting at a 90 degree angle to the scalp, six studies used a 45 degree angle (De Jesus 2011; Fitzgerald 2005; Gao 2009b; Guse 2013; Hoffman 2005; McIntosh 2004) and two at 180 degrees (Hao 2008, Liu 2008). Zhang 2010 and Zheng 2012 used the reverse side of the coil plane to the scalp. NCT00308997 used placebo stimulation, which feels similar to real rTMS but does not produce direct brain effects, Vercammen 2009a used sham designed to produce an identical sound, Wing 2012 administered sham in the single‐wing tilt position. Additionally, some studies described further sham methods which included using the same stimulation as for active TMS but with a sham coil designed to produce identical sound (Brunelin 2006; Chen 2011; Cordes 2010; Mogg 2005; Poulet 2005; Rosa 2007; Rosenberg 2012; Saba 2006a), and a sham treatment which used a magnetically non‐translucent headpiece (Schneider 2008). Bagati 2009 and Liu 2008 did not use sham but compared TMS to antipsychotics only.
5.4 Standard treatment
Two studies (Bagati 2009; Liu 2011) compared temporoparietal TMS to standard treatment, which was treatment with antipsychotics. In both trials, participants in the TMS group also received antipsychotics. We found no studies that compared prefrontal TMS to standard treatment.
6. Use of antipsychotics
In 10 studies participants in both treatment groups received first‐generation and second‐generation antipsychotics (Bagati 2009; Barr 2013; Blumberger 2012; Chen 2011; Cordes 2010; Liu 2011; McIntosh 2004; Mogg 2005; Ren 2011; Slotema 2011), although in McIntosh 2004 participants on clozapine were excluded from the trial. In 12 studies participants used second‐generation antipsychotics: in De Jesus 2011 and Rosa 2007 all participants took clozapine; in Fitzgerald 2005 a significant number in each treatment group used clozapine; in Gao 2009b and Yu 2010 participants received risperidone; and in three studies all participants used second‐generation antipsychotics apart from one participant in the TMS group who used first‐generation antipsychotics (Holi 2004; Novak 2006; Fitzgerald 2008). Six studies did not report whether antipsychotics were used in the study (Brunelin 2006; Lee 2005; NCT00308997; Wing 2012; Xu 2011; Zhang 2010) and in the remaining studies all participants received antipsychotics, but the type was not reported.
7. Outcomes
A variety of scales, used to assess clinical response and cognitive performance, are described in Appendix 1. They assessed global state, mental state, cognitive state, adverse events and quality of life.
8. Missing outcomes
No usable data were available for a number of outcomes, including adverse events, hospital and service outcomes, satisfaction with care, and economic outcomes.
Excluded studies
We excluded 58 studies. Reasons for exclusion were that 20 studies were not randomised controlled trials; one study was not randomised and the number of participants was less than five (Hoffman 1999); one study used no allocation concealment (Jandl 2006); one study included participants with depression and not schizophrenia (Hasey 2000); one study used healthy controls (NCT01620086); for 12 studies the intervention was transcranial direct current stimulation and not TMS (ACTRN12611000731998; ACTRN12612000217808; ACTRN12612001112853; Brunelin 2012; Mattai 2011; NCT00757497; NCT00870909; NCT01378078; NCT01607840; NCT01623726; Rushby 2010; Weickert 2010); for one study the intervention was an antidepressant plus fMRI and not TMS (NCT01041274); for two studies both intervention and comparison arms included TMS (NCT01595503; Slotema 2012); for one study the number of participants in each arm of the trial was less than five (Schonfeldt‐Lecuona 2004); 16 studies provided insufficient data for use (Alva 2001; Arends 2005; Benitez 2005; Cordes 2008; Daskalakis 2007; Grenier 2008; Hajak 2004; Hasan 2010; Hoffman 2000; Hoffman 2003; Jin 2003; Jin 2006; Loo 2010; Mobascher 2005; Potkin 2000; Rollnik 2000; Schneider 2001); and one study was terminated as they were unable to recruit participants (NCT00517075). We excluded Jin 2012 as data were not reported separately for temporoparietal and prefrontal TMS
Awaiting assessment
There is one study, Mohr 2006, awaiting assessment because we could not find the full article. See Characteristics of studies awaiting classification for more details.
Ongoing studies
We identified 18 ongoing studies, with 790 planned participants. One trial out of the 18 ongoing trials compares TMS with treatment as usual, and the remaining studies compare TMS with sham TMS, although one trial (NCT01370291) plans to compare both treatments with and without the use of risperidone. Three studies use high‐frequency prefrontal TMS, five use low‐frequency temporoparietal TMS, but a further nine studies use other TMS procedures (deep‐coil TMS in one study, high‐frequency temporoparietal TMS in another, and theta burst stimulation TBS in seven). See Characteristics of ongoing studies for details of each study.
Risk of bias in included studies
See Characteristics of included studies for our judgements and motivation for risk of bias for each study, Figure 2 for an overview of our judgements of risk of bias for each study and Figure 3 for an overview of percentages of low, unclear and high risk of bias for each category.
Risk of bias summary: review authors' judgements about each risk of bias item for each included study.
Risk of bias graph: review authors' judgements about each risk of bias item presented as percentages across all included studies.
Allocation
All included studies were reported as randomised. Seventeen studies adequately described the method of sequence generation (Bagati 2009; Chen 2011; De Jesus 2011; Fitzgerald 2008; Gao 2009a; Gao 2009b; Guse 2013; Hao 2008; Hoffman 2005; Liu 2008; Liu 2011; McIntosh 2004; Prikryl 2007; Slotema 2011; Yu 2010; Zhang 2010; Zheng 2012) and thus had a low risk of selection bias; the remaining studies did not provide details and were at unclear risk of selection bias. Seven studies were rated at low risk of bias as they had adequate allocation concealment (De Jesus 2011; Fitzgerald 2005; Fitzgerald 2008; Holi 2004; McIntosh 2004; Mogg 2005; Slotema 2011). However, most studies had unclear allocation concealment.
Blinding
Only seven studies adequately described the blinding of participants and personnel (Brunelin 2006; Hoffman 2005; Poulet 2005; Prikryl 2007; Rosenberg 2012; Saba 2006a; Schneider 2008) and had a low risk of performance bias, and 34 studies had unclear risk of performance bias as the method of blinding participants and personnel was not adequately described. Most studies had a low risk of detection bias as the raters were adequately blinded, but with 13 studies at unclear risk of detection bias as they did not adequately describe blinding of outcome assessment (Gao 2009a; Gao 2009c; Gao 2010; Hao 2008; Klirova 2010; Liu 2008; Liu 2011; Ren 2010; Ren 2011; Wing 2012; Xu 2011; Yu 2010; Zhang 2010).
Incomplete outcome data
Most studies had an unclear risk of attrition bias because reasons for loss to follow‐up were not consistently indicated or were unreported. Nineteen studies had a low risk of attrition bias: three studies were analysed on an intention‐to‐treat basis (Blumberger 2012; Hoffman 2005; Mogg 2005), seven studies adequately reported and dealt with attrition (Barr 2013; Chen 2011; Cordes 2010; Liu 2008; NCT00308997; Zhang 2010; Zheng 2012) and nine studies reported no losses to follow‐up (Gao 2009b; Gao 2009c; Liu 2011; McIntosh 2004; Poulet 2005; Ren 2010; Ren 2011; Xu 2011; Yu 2010). Two studies had a high risk of attrition bias. For Wing 2012, losses to follow‐up were not balanced between treatment groups, and Rosenberg 2012 had a very high (44%) attrition rate.
Selective reporting
Most studies had a low risk of reporting bias as they fully reported all stated outcomes. In five studies we considered the risk of reporting bias to be unclear (De Jesus 2011; Klirova 2010; Poulet 2005; Prikryl 2007; Wing 2012). In 14 studies we considered the risk of reporting bias to be high, as some stated outcomes were not adequately reported (Bagati 2009; Fitzgerald 2005; Fitzgerald 2008; Gao 2009a; Gao 2009c; Guse 2013; Holi 2004; Klein 1999; Liu 2011; McIntosh 2004; Novak 2006; Schneider 2008; Yu 2010; Zhang 2010). We attempted to obtain any data which were not reported in published literature by contacting the authors.
Other potential sources of bias
We rated 22 studies at low risk of bias, as we detected no other potential sources of bias. The remaining 19 had an unclear risk of bias as there was insufficient information to make a judgement.
Effects of interventions
See: Summary of findings for the main comparison TEMPOROPARIETAL TMS compared to SHAM TMS for schizophrenia; Summary of findings 2 TEMPOROPARIETAL TMS compared to STANDARD TREATMENT for schizophrenia; Summary of findings 3 PREFRONTAL TMS compared to SHAM TMS for schizophrenia; Summary of findings 4 PREFRONTAL TBS TMS compared to SHAM TMS for schizophrenia
COMPARISON 1: TEMPOROPARIETAL TMS vs SHAM TMS
Twenty trials randomised 692 participants and compared TEMPOROPARIETAL TMS (n = 399) vs SHAM TMS (n = 293) (Blumberger 2012; Brunelin 2006; De Jesus 2011; Fitzgerald 2005; Gao 2009a; Gao 2010; Hao 2008; Hoffman 2005; Klirova 2010; Lee 2005; Liu 2008; McIntosh 2004; NCT00308997; Rosa 2007; Rosenberg 2012; Saba 2006a; Slotema 2011; Vercammen 2009a; Xu 2011; Yu 2010).
1.1 Global state
a. Clinical improvement (CGI)
One study (Gao 2009a) found that the number of participants with a clinical improvement in global state did not differ between temporoparietal TMS and sham TMS when measured on the CGI scale; however they did not report the response criteria used to define clinical improvement (Analysis 1.1; 46 participants).
b. Average scores for clinical improvement (CGI, high = poor)
Seven studies reported global state measured on the CGI scale and found a clear difference in favour of temporoparietal TMS (7 RCTs, n = 224, MD ‐0.50, 95% CI ‐0.76 to ‐0.23, Analysis 1.2).
1.2 Mental state
a. General
i. Clinical improvement (PANSS > 30% reduction)
Blumberger 2012 reported clinical improvement in mental state, defined as more than a 30% reduction in total PANSS score; the proportion of participants that had a clinical improvement in mental state did not differ between the treatment groups (Analysis 1.3; 51 participants).
ii. Average total score (various scales)
Mental state was measured on the BPRS by De Jesus 2011 (17 participants), which found no clear difference in scores between treatment groups. In contrast, total PANSS scores were clearly lower in the temporoparietal TMS group than the sham TMS group (5 RCTs, n = 127, MD ‐6.09, 95% CI ‐10.95 to ‐1.22, Analysis 1.4).
iii. Average general psychopathology score (PANSS general)
Four studies provided data regarding general psychopathology measured on the PANSS general subscale. There was no significant difference in scores between temporoparietal TMS and sham TMS (Analysis 1.5; 87 participants).
b. Specific
i. Average depression score (various scales)
Hao 2008 found that participants showed significantly less depression when measured on the SDS (1 RCT, n = 25, MD ‐5.59, 95% CI ‐11.57 to 0.39, Analysis 1.6) , but results were equivocal when measured on the HAMD by the same small study. De Jesus 2011 also reported data for depression and excitement factor on the BPRS, but these data were skewed so we have not presented them in analyses (see Table 2).
| Study | Outcome | TMS Mean | TMS SD | TMS N | Sham TMS Mean | Sham TMS SD | Sham TMS N |
| Mental state: specific – BPRS depressive factor (high = poor) | 2.25 | 2.18 | 8 | 3.56 | 3.24 | 9 | |
| Mental state: specific – BPRS excitement factor (high = poor) | 1.25 | 1.28 | 8 | 3.89 | 4.79 | 9 | |
| Mental state: Specific ‐ positive symptoms (SAPS, high = poor) | 51.2 | 13.5 | 5 | 47.8 | 25.2 | 5 | |
| Mental state: Specific ‐ hallucinations (AHRS, high = poor) | 14.6 | 12.1 | 5 | 20.8 | 3.4 | 5 | |
| Mental state: Specific ‐ positive symptoms (SAPS, high = poor) | 26 | 20 | 10 | 37 | 16 | 10 | |
| Mental state: Specific ‐ negative symptoms (SANS, high = poor) | 32 | 27 | 10 | 39 | 23 | 10 | |
| Cognitive state: CPT false items | 1.94 | 2.04 | 18 | 1.41 | 2.12 | 17 | |
| Cognitive state: CPT missed items | 6.28 | 4.5 | 18 | 7.59 | 6.68 | 17 |
AHRS ‐ Auditory Hallucination Rating Scale
BPRS ‐ Brief Psychiatric Rating Scale
CPT ‐ Continuous Performance Test
SANS ‐ Scale for Assessment of Negative Symptoms
SAPS ‐ Scale for the Assessment of Positive Symptoms
ii. Hallucinations ‐ clinical improvement (various scales)
Significantly more participants that received temporoparietal TMS showed a clinical improvement in hallucinations when defined as an HCS score of 5 or less (3 RCTs, n = 133, RR 2.26, 95% CI 1.18 to 4.35) or more than a 30% decrease on the AHRS (3 RCTs, n = 120, RR 2.99, 95% CI 1.12 to 7.98, Analysis 1.7). However, AHRS pooled data showed moderate heterogeneity (I² = 55%) and when we applied the random‐effects model the results became non‐significant. Rosa 2007 reported “Reality” and “Attentional Salience” scores from the AHRS in figures, which showed a significant group effect (P = 0.0493 and P = 0.0360, respectively).
We found no clear difference for clinical improvement in hallucinations when defined as improvement of one or more points on the PANSS hallucination item score, or more than a 30% reduction on the PSYRATS score.
iii. Average hallucinations score (various scales)
Hallucinations scores were significantly lower in the temporoparietal TMS when measured on the HCS (3 RCTs, n = 162, MD ‐1.64, 95% CI ‐2.80 to ‐0.48) and by the PANSS hallucination item (4 RCTs, n = 125, MD ‐1.01, 95% CI ‐1.97 to ‐0.04, Analysis 1.8). However, the PANSS hallucination item data were highly heterogenous (I² = 81%), with no obvious clinical or methodological reason for the heterogeneity. Removal of the outlying study, Gao 2010, reduced the heterogeneity (I² = 30%).
In contrast, when hallucinations were measured using AVH‐related items from the PSYRATS and the AHRS, there was no significant difference in hallucination scores between treatment groups. However, the latter showed high levels of heterogeneity (I² = 62%), which we could not explain by differences in the treatment as all used low frequency (1 Hz). Furthermore, when we removed Rosenberg 2012, which used deep temporoparietal TMS, the heterogeneity was unchanged. However, when we removed change data from the analysis (NCT00308997; Poulet 2005) the heterogeneity was reduced (I² = 20%).
Poulet 2005 also reported endpoint data for hallucinations on the AHRS, but these data were skewed so we have not presented them in analyses (see Table 2).
iv. Average negative symptom scores (various scales)
Negative symptoms were measured using the BPRS, PANSS negative and SANS scales. We found no significant difference in scores on the BPRS and PANSS; however one small study, Hao 2008, which used high‐frequency temporoparietal TMS (10 Hz) showed a significant difference favouring temporoparietal TMS (1 RCT, n = 25, MD ‐23.58, 95% CI ‐37.06 to ‐10.1, Analysis 1.9). Rosenberg 2012 also reported data for negative symptoms on the SANS, but these data were skewed so we have not presented them in analyses (see Table 2).
v. Positive symptoms ‐ clinical improvement (PANSS > 30% reduction)
Blumberger 2012 found no difference in clinical improvement of positive symptoms, which was defined as more than a 30% reduction in PANSS positive subscale score (Analysis 1.10; 51 participants).
vi. Average positive symptom score (various scales)
Positive symptom scores were significantly lower in the temporoparietal TMS group than in the sham TMS group when measured on the PANSS positive subscale (11 RCTs, n = 333, MD ‐2.14, 95% CI ‐3.15 to ‐1.14, Analysis 1.11), but not significantly different when measured on the BPRS in one study (De Jesus 2011; 17 participants) or the SAPS used by Brunelin 2006 and Hao 2008. Poulet 2005 and Rosenberg 2012 also reported data for positive symptoms on the SAPS, but these data were skewed so we have not presented them in analyses (see Table 2).
1.3 Cognitive state
Cognitive state was reported in three studies (Hoffman 2005; Liu 2008; Xu 2011) using 39 different measures. These data are reported in Table 3. Results were equivocal for all measures. Xu 2011 also reported cognitive data on the CPT, but these data were skewed so we did not present them in Table 3 (see Table 2).
| Outcome | Change / endpoint data | Study | TMS | Sham TMS | Mean difference [95% CI] | ||||
| Mean | SD | N | Mean | SD | N | ||||
| Animal naming | Change | ‐0.77 | 4.41 | 26 | 0.9 | 4.17 | 21 | ‐1.67 [‐4.13 to 0.79] | |
| CPT reaction time (ms) | Endpoint | 926.22 | 126.2 | 18 | 959 | 109.35 | 17 | ‐32.78 [‐110.89 to 45.33] | |
| Controlled oral word association | Change | 2.57 | 7.07 | 26 | 2.53 | 0.91 | 21 | 0.04 [‐2.71 to 2.79] | |
| CVLT 1 score | Change | 0.88 | 6.61 | 26 | ‐0.19 | 1.69 | 21 | 1.07 [‐1.57 to 3.71] | |
| CVLT B score | Change | 0.15 | 1.76 | 26 | 0.48 | 3.1 | 21 | ‐0.33 [‐1.82 to 1.16] | |
| CVLT Long‐delay free recall | Change | ‐1.69 | 2.28 | 26 | ‐1.48 | 3.1 | 21 | ‐0.21 [‐1.80 to 1.38] | |
| CVLT Recognition discriminative ability | Change | ‐0.007 | 0.08 | 26 | ‐0.014 | 0.088 | 21 | 0.01 [‐0.04 to 0.06] | |
| CVLT Short‐delay free recall | Change | ‐0.92 | 2.15 | 26 | ‐0.71 | 3 | 21 | ‐0.21 [‐1.74 to 1.32] | |
| CVLT1‐5 Total score | Change | ‐3.42 | 7.08 | 26 | ‐3.14 | 8.21 | 21 | ‐0.28 [‐4.72 to 4.16] | |
| Digit recall (distraction) | Change | 0.61 | 3.93 | 26 | ‐0.9 | 4.77 | 21 | 1.51 [‐1.03 to 4.05] | |
| Digit recall (non‐distraction) | Change | ‐0.12 | 4.66 | 26 | 1.19 | 4.08 | 21 | ‐1.31 [‐3.81 to 1.19] | |
| Digit symbol | Change | 3.15 | 7.76 | 26 | 2.95 | 7.72 | 21 | 0.20 [‐4.25 to 4.65] | |
| Grooved pegboard, dominant | Change | 4.65 | 15.1 | 26 | 5.57 | 48.8 | 21 | ‐0.92 [‐22.58 to 20.74] | |
| Grooved pegboard, nondominant | Change | 6.46 | 15.5 | 26 | 12 | 31.5 | 21 | ‐5.54 [‐20.27 to 9.19] | |
| Temporal orientation | Change | ‐0.154 | 1.82 | 26 | ‐0.35 | 2.39 | 21 | 0.20 [‐1.04 to 1.43] | |
| Trail Making A | Change | 2.58 | 12.6 | 26 | ‐0.42 | 8.23 | 21 | 3.00 [‐2.99 to 8.99] | |
| Trail Making B | Change | 19.5 | 48.3 | 26 | 25.3 | 50.4 | 21 | ‐5.80 [‐34.25 to 22.65] | |
| WCST completed categories | Endpoint | 2.17 | 2.23 | 18 | 2.82 | 2.32 | 17 | ‐0.65 [‐2.16 to 0.86] | |
| WCST completed categories | Endpoint | 5.3 | 1 | 11 | 4.5 | 1.4 | 10 | 0.80 [‐0.25 to 1.85] | |
| WCST conceptualisation level | Endpoint | 61 | 24.13 | 18 | 64.12 | 24.93 | 17 | ‐3.12 [‐19.39 to 13.15] | |
| WCST CR | Endpoint | 54 | 9 | 11 | 49 | 11 | 10 | 5.00 [‐3.65 to 13.65] | |
| WCST FM | Endpoint | 13.6 | 7.3 | 11 | 9.8 | 11.2 | 10 | 3.80 [‐4.37 to 11.97] | |
| WCST NPE | Endpoint | 17.2 | 7.6 | 11 | 20.9 | 5.6 | 10 | ‐3.70 [‐9.38 to 1.98] | |
| WCST PCLR | Endpoint | 60.5 | 19.2 | 11 | 45.9 | 18.6 | 10 | 14.60 [‐1.58 to 30.78] | |
| WCST PE | Endpoint | 45.3 | 23.6 | 11 | 50.8 | 22.3 | 10 | ‐5.50 [‐25.14 to 14.14] | |
| WCST PNPE | Endpoint | 31.3 | 13.8 | 11 | 32.4 | 14.5 | 10 | ‐1.10 [‐13.24 to 11.04] | |
| WCST PPE | Endpoint | 68.7 | 13.8 | 11 | 67.6 | 14.6 | 10 | 1.10 [‐11.08 to 13.28] | |
| WCST PR | Endpoint | 41 | 13.3 | 11 | 34.2 | 15 | 10 | 6.80 [‐5.37 to 18.97] | |
| WCST Ra | Endpoint | 122.67 | 15.18 | 18 | 126.06 | 5.02 | 17 | ‐3.39 [‐10.80 to 4.02] | |
| WCST Re | Endpoint | 56.11 | 22.99 | 18 | 53.88 | 16.14 | 17 | 2.23 [‐10.87 to 15.33] | |
| WCST Rp | Endpoint | 45.72 | 20.18 | 18 | 60.12 | 19.23 | 17 | ‐14.40 [‐27.46 to ‐1.34] | |
| WCST TA | Endpoint | 117 | 18 | 11 | 121 | 10 | 10 | ‐4.00 [‐16.31 to 8.31] | |
| WCST TCFC | Endpoint | 21.7 | 14 | 11 | 29.0 | 13.4 | 10 | ‐7.30 [‐19.02 to 4.42] | |
| WCST TE | Endpoint | 63 | 24 | 11 | 72 | 20 | 10 | ‐9.00 [‐27.84 to 9.84] | |
| WCST time (sec) | Endpoint | 405 | 174 | 11 | 411 | 177 | 10 | ‐6.00 [‐156.36 to 144.36] | |
| WCST selective error rate (%) | Endpoint | 51.6 | 15.3 | 11 | 58.4 | 12.3 | 10 | ‐6.80 [‐18.63 to 5.03] | |
| WCST correct thinking time (sec) | Endpoint | 172 | 67 | 11 | 160 | 96 | 10 | 12.00 [‐59.47 to 83.47] | |
| WCST error thinking time (sec) | Endpoint | 233 | 128 | 11 | 251 | 100 | 10 | ‐18.00 [‐115.79 to 79.79] | |
| WRAT‐R | Change | 0.19 | 2.54 | 26 | 0.33 | 2.81 | 21 | 6.80 [‐5.37 to 18.97] | |
CPT ‐ Continuous performance test
CVLT ‐ California verbal learning test
WCST ‐ Wisconsin card sorting test
WRAT‐R ‐ wide range achievement test – reading
1.4 Adverse effects
a. General
i. Serious
NCT00308997 and Vercammen 2009a reported that there were no serious adverse events in either treatment group (Analysis 1.12; 130 participants).
ii. Leaving the study early
The number of participants leaving the study early did not differ significantly between treatment groups (Analysis 1.13; 8 studies, 320 participants).
b. Specific
Participants receiving temporoparietal TMS clearly experienced more headaches (10 RCTs, n = 392, RR 2.65, 95% CI 1.56 to 4.50) and jaw and facial contraction (2 RCTs, n = 70, RR 8.32, 95% CI 1.13 to 61.17, Analysis 1.14) than those receiving sham TMS. Other adverse events ‐ concentration problems, earache, lightheadedness/dizziness, mild memory impairment/amnesia, restless legs, somatic discomfort, tingling sensation in the arm, worsening hallucinations/audible thoughts ‐ were not clearly different between treatment groups.
1.5 Quality of life
a. Average score (Q‐LES‐Q, low = poor)
Rosenberg 2012 measured quality of life on the Q‐LES‐Q and found no clear difference between deep temporoparietal TMS and sham TMS (Analysis 1.15; 20 participants).
COMPARISON 2: TEMPOROPARIETAL TMS vs STANDARD TREATMENT
Two trials randomised 140 participants and compared TEMPOROPARIETAL TMS (n = 70) versus STANDARD TREATMENT (n = 70) (Bagati 2009; Liu 2011). In both studies the participants received first‐ and second‐generation antipsychotics in both treatment groups.
2.1 Global state
a. Clinical improvement (CGI ≤2)
Liu 2011 found that there was no clear difference in the number of participants experiencing clinical improvement when temporoparietal TMS and antipsychotics were compared to antipsychotic treatment alone (Analysis 2.1; 100 participants).
2.2 Mental state
a. Average hallucinations score (AHRS)
Bagati 2009 reported data for hallucinations on the AHRS, but these data were skewed so we have not presented them in analyses (see Table 4).
| Study | Outcome | TMS Mean | TMS SD | TMS N | Sham TMS Mean | Sham TMS SD | Sham TMS N |
| Mental state: Specific ‐ hallucinations (AHRS, high = poor) | 6.7 | 8.64 | 20 | 27.95 | 7.51 | 20 |
AHRS ‐ Auditory Hallucination Rating Scale
2.3 Adverse effects
a. General ‐ leaving the study early
The number of participants leaving the study early was not clearly different between temporoparietal TMS and antipsychotics alone (Analysis 2.2; 140 participants).
COMPARISON 3: PREFRONTAL TMS vs SHAM TMS
Seventeen trials randomised 502 participants and compared PREFRONTAL TMS (n = 266) versus SHAM TMS (n = 236) (Barr 2013; Cordes 2010; Fitzgerald 2008; Gao 2009b; Gao 2009c; Guse 2013; Holi 2004; Klein 1999; Mogg 2005; Novak 2006; Poulet 2005; Prikryl 2007; Ren 2010; Ren 2011; Schneider 2008; Wing 2012; Zheng 2012).
3.1 Global state
a. Average scores (various scales)
Three small studies (Guse 2013; Holi 2004; Klein 1999) measured global state on the CGI, CGI‐S, GAF and SCL‐90, none of which showed a significant effect between prefrontal TMS and sham TMS (Analysis 3.1; 85 participants).
3.2 Mental state
a. General
i. Clinical improvement (> 20% decrease in total PANSS score)
Results from one small trial (Holi 2004) show that more participants in the prefrontal TMS group had a clinical improvement in mental state than those that received sham TMS (1 RCT, n = 22, RR 0.14, 95% CI 0.02 to 0.98, Analysis 3.2).
ii. Average total score (various scales)
Mental state was measured using the BPRS and PANSS scales. We found no clear difference in participants' mental state between treatment groups on either scale (Analysis 3.3; 219 participants). However, the pooled data for the PANSS scale were heterogeneous (I² = 68%). When we removed the low‐frequency trial (Ren 2010), the heterogeneity remained. Removing outlying trials, Fitzgerald 2008 and Gao 2009b, eliminated the heterogeneity, but the studies account for more than 40% of the weight for this outcome and we therefore did not pool the data.
iii. Average general psychopathology score (PANSS, high=poor)
Six studies reported data on general psychopathology of participants, measured on the PANSS scale. The pooled data were highly heterogenous (I² = 81%) and removal of the two low‐frequency trials (Klein 1999; Ren 2010) did not reduce the heterogeneity. Removal of the outlying trials, Gao 2009b and Klein 1999, eliminated the heterogeneity, but as these trials accounted for 36% of the weighting, we did not pool the data (Analysis 3.4; 199 participants).
b. Specific
i. Average depression score (various scales)
Depression was reported on four scales by four different studies. There were no significant differences on the HDRS and SCL‐90 DEP subscale, whereas, when measured on the HAMD‐17 (1 RCT, n = 43, MD ‐2.40, 95% CI ‐3.88 to ‐0.92) and MADRS (1 RCT, n = 22, MD ‐4.36 95% CI ‐7.05 to ‐1.67), prefrontal TMS was efficacious when compared to sham TMS (Analysis 3.5). Barr 2013 and Fitzgerald 2008 also reported data for depression on the Calgary depression scale (CDS) and Calgary depression rating scale (CDRS) respectively, but as these data are skewed we have not presented them in analyses (see Table 5).
| Study | Outcome | TMS Mean | TMS SD | TMS N | Sham TMS Mean | Sham TMS SD | Sham TMS N |
| Mental state: Specific ‐ negative symptoms (PANSS, high = poor) | 26.15 | 13.45 | 13 | 31.42 | 13.19 | 12 | |
| Mental state: Specific – depressive symptoms (CDS, high = poor ) | 2.38 | 2.06 | 13 | 1.67 | 1.92 | 12 | |
| Mental state: Specific ‐ positive symptoms (PANSS, high = poor) (LOCF) | 10.8 | 7.0 | 10 | 7.3 | 2.9 | 10 | |
| Mental state: Depression (CDRS, high = poor) (LOCF) | 7.2 | 5.9 | 10 | 3.5 | 3.8 | 10 |
CDRS ‐ Calgary depression rating scale
CDS ‐ Calgary depression scale
PANSS ‐ positive and negative symptoms scale
LOCF ‐ last observation carried forward
CDRS ‐ Calgary depression rating scale
CDS ‐ Calgary depression scale
ii. Average hallucinations score (PANSS)
Ren 2010 found no difference in hallucinations between treatment groups (1 RCT, n = 25, MD ‐0.68, 95% CI ‐1.68 to 0.32, Analysis 3.6).
iii. Negative symptoms ‐ clinical improvement (> 20% decrease in PANSS negative)
One small study (Novak 2006) found no difference in the number of participants that experienced a clinical improvement in negative symptoms (Analysis 3.7; 16 participants).
iv. Average negative symptom score (various scales)
Pooled data for 10 studies that reported negative symptoms on the PANSS positive subscale were highly heterogeneous. Removing the low‐frequency studies (Fitzgerald 2008; Klein 1999) did not reduce the heterogeneity. Removal of the outlying trials (Gao 2009b; Gao 2009c) reduced the heterogeneity (I² = 16%), and results show no significant difference between treatment groups. When measured on the SANS, three small studies found that participants receiving prefrontal TMS had a significant improvement compared to sham TMS (3 RCTs, n = 71, MD ‐12.68, 95% CI ‐18.60 to ‐6.77, Analysis 3.8). Barr 2013 also reported data for negative symptoms on the PANSS, but these data were skewed so we have not presented them in analyses (see Table 5).
v. Average positive symptom score (various scales)
Positive symptoms were not significantly different between treatment groups for 10 studies (279 participants) that used the PANSS positive subscale and one small study (Prikryl 2007; 22 participants) on the SAPS (Analysis 3.9). Fitzgerald 2008 also reported data for positive symptoms on the PANSS, but these data were skewed so we have not presented them in analyses (see Table 5).
vi. Average psychotism score (SCL‐90 PSY)
Holi 2004 also found no difference in psychotism measured on the SCL‐90 PSY subscale between prefrontal TMS and sham TMS (Analysis 3.10; 22 participants).
3.3 Cognitive state
Cognitive effects were reported in four studies (Guse 2013; Mogg 2005; Novak 2006; Zheng 2012), using 19 different measures. These data are reported in Table 6. Results were equivocal for most of the outcome measures, with limited evidence to suggest a beneficial effect of TMS for five cognitive test scores. One study (Mogg 2005) reported significantly increased cognitive test scores on average in the TMS arm compared with the control arm for four outcomes: Hopkins verbal learning test (HVLT)‐delayed recall (after two weeks follow‐up), controlled oral word association test (COWAT) (two weeks after TMS), and the Stroop test (within 24 hours of treatment and at two weeks follow‐up) . A second study (Guse 2013) reported significantly increased scores for Wisconsin card sorting test (WCST) categories (for people with WCST categories pre‐treatment median ≤ 4). More trials are needed to confirm or refute the beneficial effects of these cognitive test outcomes.
| Outcome | Change / endpoint | Study | TMS | Sham TMS | Mean difference [95% CI] | ||||
| Mean | SD | N | Mean | SD | N | ||||
| AVLT (low = poor) | Endpoint | 45.6 | 6.8 | 8 | 44.9 | 8 | 8 | 0.70 [‐6.58 to 7.98] | |
| COWAT (within 24 hours of TMS) | Endpoint | 11.6 | 5.3 | 8 | 10.9 | 5.0 | 9 | 0.70 [‐4.22 to 5.62] | |
| COWAT (2 weeks after TMS) | Endpoint | 14.2 | 5.7 | 8 | 9.1 | 2.7 | 9 | 5.10 [0.77 to 9.43] | |
| Digit span test | Endpoint | 10.5 | 3.5763 | 38 | 9.5 | 4 | 17 | 1.00 [‐15.70 to 17.70] | |
| Grooved pegboard (seconds to completion) (within 24 hours of TMS) | Endpoint | 117.1 | 32.0 | 8 | 108.6 | 41.2 | 9 | 8.50 [‐26.37 to 43.37] | |
| Grooved pegboard (seconds to completion) (2 weeks after TMS) | Endpoint | 109 | 29.5 | 8 | 98.5 | 16 | 9 | 10.50 [‐12.46 to 33.46] | |
| HVLT‐ delayed recall (within 24 hours of TMS) | Endpoint | 4.4 | 2.3 | 8 | 4.4 | 1.1 | 9 | 0.00 [‐1.75 to 1.75] | |
| HVLT‐ delayed recall (2 weeks after TMS) | Endpoint | 5.4 | 2.7 | 8 | 3.3 | 1.0 | 9 | 2.10 [0.12 to 4.08] | |
| HVLT‐ immediate recall (within 24 hours of TMS) | Endpoint | 6.3 | 2.0 | 8 | 5.6 | 1.1 | 9 | 0.70 [‐0.86 to 2.26] | |
| HVLT‐ immediate recall (2 weeks after TMS) | Endpoint | 7.4 | 2.8 | 8 | 5 | 0.8 | 9 | 2.40 [0.39 to 4.41] | |
| Stroop test (within 24 hours of TMS) | Endpoint | 77.4 | 20.3 | 8 | 51.4 | 14.9 | 9 | 26.00 [8.89 to 43.11] | |
| Stroop test (2 weeks after TMS) | Endpoint | 88.2 | 12.3 | 8 | 60.8 | 6.4 | 9 | 27.40 [17.91 to 36.89] | |
| Trail making test A | Change | 0.64 | 15.08 | 14 | ‐11.92 | 29.27 | 12 | 12.56 [‐5.79 to 30.91] | |
| Trail making test B | Change | ‐0.54 | 28.41 | 13 | ‐5.64 | 20.31 | 11 | 5.10 [‐14.46 to 24.66] | |
| Verbal fluency test (high = poor) | Endpoint | 24.2 | 9.2542 | 38 | 24.5 | 6.6 | 17 | ‐0.30 [‐4.60 to 4.00] | |
| WCST categories | Change | 1.58 | 22.2 | 12 | ‐0.27 | 1.95 | 11 | 1.85 [‐10.76 to 14.46] | |
| WCST categories for participants with WCST categories pre < median (= 4) | Change | 3.33 | 2.58 | 6 | 0.4 | 2.07 | 5 | 2.93 [0.18 to 5.68] | |
| WCST perseverative answers | Change | ‐9 | 11.65 | 12 | ‐19.18 | 27.76 | 11 | 10.18 [‐7.50 to 27.86] | |
| WCST perseverative mistakes | Change | ‐8.17 | 9.81 | 12 | ‐11.27 | 17.51 | 11 | 3.10 [‐8.64 to 14.84] | |
AVTL ‐ auditory verbal learning test
COWAT ‐ controlled oral word association test
HVLT ‐ Hopkins verbal learning test
WCST ‐ Wisconsin card sorting test
3.4 Adverse effects
a. General
i. Leaving the study early
The number of participants leaving the study was reported in eight studies and did not differ significantly between treatment groups (8 RCTs, n = 174, RR 1.19, 95% CI 0.56 to 2.50, Analysis 3.12).
b. Specific
i. Various
Participants in the prefrontal TMS group experienced more headaches than those in the sham TMS group (6 RCTs, n = 164, RR 2.77, 95% CI 1.22 to 6.26), and more TMS‐related site discomfort or pain (2 RCTs, n = 42, RR 8.33, 95% CI 1.68 to 41.27, Analysis 3.13). Cordes 2010 reported no adverse events measured on the UKU side effect rating scale and Klein 1999 reported no cognitive difficulties in either treatment group. Klein 1999 also found no significant difference in facial twitching and worsening of pre‐existing akathisia and OCD.
ii. Average scores (CSSES)
Mogg 2005 measured subjective side effects and cognitive complaints on the CSSES and found no significant differences between prefrontal TMS and sham TMS (Analysis 3.14; 17 participants).
COMPARISON 4: PREFRONTAL TBS TMS vs SHAM TMS
Three trials randomised 115 participants and compared PREFRONTALTBS TMS (n = 59) versus SHAM TMS (n = 56) (Chen 2011; Zhang 2010; Zheng 2012).
4.1 Global state
a. Clinical improvement
Zhang 2010 found no difference in the number of participants showing a clinical improvement in global state between prefrontal TBS TMS and sham TMS (Analysis 4.1; 27 participants).
4.2 Mental state
a. General
i. Average total score (PANSS, high = poor)
Three studies (Chen 2011; Zhang 2010; Zheng 2012) reported data for mental state on the PANSS scale and found that prefrontal TBS TMS was efficacious when compared to sham TMS (3 RCTs, n = 108, MD ‐5.71, 95% CI ‐9.32 to ‐2.10, Analysis 4.2).
ii. Average general psychopathology score (PANSS, high = poor)
General psychopathology was also significantly better in the prefrontal TBS TMS group (3 RCTs, n = 108, MD ‐2.47, 95% CI ‐4.21 to ‐0.73, Analysis 4.3).
b. Specific
i. Average negative symptom score (PANSS, high = poor)
Negative symptoms were significantly lower in the prefrontal TBS TMS group than in the sham TMS group when measured on the PANSS (3 RCTs, n = 108, MD ‐2.67, 95% CI ‐4.25 to ‐1.09) and the SANS (1 RCT, n = 27, MD ‐11.55, 95% CI ‐21.90 to ‐1.2, Analysis 4.4).
ii. Average positive symptom score (PANSS, high = poor)
Positive symptoms were not significantly different between treatment groups (Analysis 4.5; 108 participants).
4.3 Cognitive state
a. Average scores on various measures
We found no difference in cognitive state between treatment groups when measured using the digit span test and the verbal fluency test in one small study (Zheng 2012) (Analysis 4.6; 39 participants).
4.4 Adverse effects
a. General ‐ Leaving the study early
The number of participants leaving the study early did not differ between the treatment groups (Analysis 4.7; 2 RCTs, 76 participants).
b. Specific
Participants did not experience significantly different numbers of adverse events (headaches or sleep disorder) between prefrontal TBS TMS and sham TMS (Analysis 4.8; 1 RCT, 27 participants).
Unusable data
Jin 2012 reported data for clinical improvement combined for the two TMS groups (frontal and parietal) in the study, and so could not be added to any of the comparisons on the analyses. Clinical improvement was defined as at least a 30% improvement in PANSS score; 17 of 41 patients responded to the TMS (42%), whereas three of 24 responded to sham TMS (12%).
Sensitivity analysis
There were no losses to follow‐up for the outcome 'clinical improvement in global state' for temporoparietal TMS compared to sham TMS or standard treatment, and no studies reported on this outcome when prefrontal TMS was compared with sham TMS. For prefrontal TBS TMS versus sham TMS there were no differences when completer‐only data were compared with all randomised in an intention‐to‐treat analysis.
Discusión
En general la calidad de las pruebas se valoró como muy baja según las tablas "Resumen de los hallazgos" (Resumen de los hallazgos para la comparación principal; Resumen de los hallazgos 2; Resumen de los hallazgos 3; Resumen de los hallazgos 4).
Resumen de los resultados principales
COMPARACIÓN 1: EMT TEMPOROPARIETAL VERSUS SIMULADA
Pruebas de muy baja calidad de un ensayo pequeño no mostraron evidencias de un efecto de la EMT temporoparietal comparada con la EMT simulada para mejorar clínicamente el estado global. Sin embargo, hay algunas pruebas de muy baja calidad que indican que las puntuaciones del estado global en la escala CGI son superiores con la EMT temporoparietal. También hay pruebas de muy baja calidad de la escala PANSS de que la EMT temporoparietal es superior a la EMT simulada para mejorar el estado mental. Aunque puede haber algunos efectos beneficiosos con la EMT sobre la EMT simulada, la importancia clínica de algunos de los datos derivados de las escalas es incierta. Pruebas de muy baja calidad muestran que la EMT temporoparietal no afecta el estado cognitivo; sin embargo, los participantes que recibieron EMT temporoparietal presentaron más cefaleas que los del grupo de EMT simulada. No más participante abandonaron temprano el estudio en el grupo EMT temporoparietal que en el grupo EMT simulada pero, nuevamente, son pruebas de muy baja calidad. Ningún estudio informó si los participantes estaban satisfechos con la atención.
COMPARACIÓN 2: EMT TEMPOROPARIETAL VERSUS ATENCIÓN ESTÁNDAR
Pruebas limitadas de baja calidad muestran que la EMT temporoparietal no es superior al tratamiento estándar (antipsicóticos de primera y segunda generación) para mejorar clínicamente el estado global, y el número de participantes que abandonaron temprano el estudio no difirió entre la EMT temporoparietal y el tratamiento estándar. Ningún estudio informó sobre el estado mental y el estado cognitivo de los participantes, si presentaron efectos adversos y si estaban satisfechos con su atención.
COMPARACIÓN 3: EMT PREFRONTAL VERSUS SIMULADA
No se encontraron pruebas de que la EMT prefrontal es superior a la EMT simulada para mejorar el estado global, el estado mental y el estado cognitivo, aunque la calidad de las pruebas es muy baja. La EMT prefrontal no causa más cefaleas que la EMT simulada, y el número de participantes que abandonaron temprano el estudio no difirió entre los grupos de tratamiento, pero nuevamente las pruebas son de muy baja calidad. Ningún estudio informó si los participantes estaban satisfechos con la atención.
COMPARACIÓN 4: EMT CON ETS PREFRONTAL VERSUS SIMULADA
La EMT con ETS prefrontal no es superior a la EMT simulada para mejorar el estado global y el estado cognitivo pero hay algunas pruebas de que mejora el estado mental, aunque son de muy baja calidad. La EMT con ETS prefrontal no provoca que los participantes presenten más cefaleas o abandonen temprano el estudio que la EMT simulada pero, nuevamente, las pruebas son de muy baja calidad. Ningún estudio informó si los participantes estaban satisfechos con la atención.
Compleción y aplicabilidad general de las pruebas
1. Duración
Los estudios informaron diferencias sustanciales en la duración de los ensayos, que varió de cuatro días (McIntosh 2004) hasta 10 semanas (Wing 2012). Por lo tanto, este tema es potencialmente problemático para la comparación, y se debe tener cuidado con respecto a cualquier conclusión. La diferencia en la duración del estudio puede surgir de la naturaleza de las muestras poblacionales en cuanto a las altas tasas de desgaste asociadas. La falta de consistencia en los estudios con respecto a la duración del estudio también puede reflejar el aspecto novedoso de la intervención y la falta de un procedimiento estandarizado. Los datos no se estratificaron según los diferentes períodos especificados en Tipos de medidas de resultado en el protocolo, pues no hubo suficientes datos (ver Diferencias entre el protocolo y la revisión).
2. Participantes
Los participantes se clasificaron de manera consistente como con esquizofrenia o con trastorno esquizoafectivo, y la mayoría de los estudios utilizó un diagnóstico según DMS‐IV. Prikryl 2007 informó la CIE‐10 y Schneider 2008 informó el uso de ambos diagnósticos de esquizofrenia al menos un año antes de la hospitalización. Por lo tanto, los grupos de las muestras incluidas para la revisión coincidieron bien.
3. Condición control
Se informó una variedad amplia de técnicas de EMT simulada entre los estudios incluidos. Aunque en su mayoría los estudios informaron el uso de la misma estimulación que para la EMT activa, las descripciones adicionales de este procedimiento variaron, con el borde del espiral que descansaba en un ángulo de 45 grados, un ángulo de 90 grados, con una banda en contacto o con ambas bandas en contacto. Además, las descripciones adicionales incluyeron un espiral simulado que producía sonidos idénticos a los de la EMT activa y un espiral simulado con un cabezal magnéticamente no translúcido. Por lo tanto, es difícil establecer una comparación entre los resultados e interpretar los hallazgos.
Sorprendentemente, hay información muy escasa sobre la EMT comparada con otros tratamientos para la esquizofrenia. Ningún estudio comparó la EMT con otros métodos físicos de tratamiento como el tratamiento electroconvulsivo (TEC) y dos estudios (de los 41 incluidos) compararon EMT con tratamiento estándar. El tratamiento estándar en estos estudios (Bagati 2009; Liu 2011) fue con antipsicóticos, aunque los que recibieron EMT también recibieron dichos fármacos.
Un ensayo de los 18 en curso compara EMT con tratamiento habitual y los estudios restantes comparan EMT con EMT simulada, aunque un ensayo (NCT01370291) se propone comparar ambos tratamientos con y sin el uso de risperidona. Lo anterior indica que la base de pruebas para la EMT aún se estudia contra la EMT simulada, y que todavía no se avizoran comparaciones con tratamientos activos.
4. Intervención
La intervención con EMT activa en las condiciones prefrontales y temporoparietales varió de manera significativa entre los estudios en cuanto a la intensidad de la estimulación, la duración de la estimulación y la ubicación de la EMT. Los estudios que realizaron EMT prefrontal informaron las mayores variaciones. La intensidad de la estimulación incluyó rangos de 1 Hz, 10% por encima del umbral a 20 Hz en el umbral motor del 80%, y tres estudios utilizaron ETS de 50 Hz. La duración de la estimulación en los estudios de EMT prefrontal variaron de dos series de un minuto con una brecha de tres minutos (Klein 1999), a 40 series de 2,5 segundos con una brecha de 30 segundos (Novak 2006) con algunas variaciones diferentes entre los estudios. La ubicación de la estimulación de la EMT prefrontal también difirió, con informes de EMT prefrontal izquierda, EMT de la corteza prefrontal dorsolateral izquierda, EMT prefrontal derecha y EMT prefrontal bilateral. En los estudios que utilizaron EMT temporoparietal, hubo cierta consistencia en el hecho de que todos los estudios excepto dos informaron el uso de EMT de 1 Hz. Sin embargo, el nivel del umbral motor varió, con informes del 20% por debajo del umbral motor hasta un umbral motor del 100%. Al igual que en los estudios de EMT prefrontal, en el caso de la EMT temporoparietal se informó una variedad amplia en la duración de la estimulación, que varió de cinco sesiones de un minuto con brechas de un minuto (Saba 2006a) a dos sesiones de 20 minutos cada una al día(Vercammen 2009a).También hubo más consistencia en los estudios de EMT temporoparietal en cuanto a la ubicación, ya que todos los estudios excepto uno informaron EMT temporoparietal izquierda, con la excepción de Lee 2005 que informó que también se utilizó la EMT temporoparietal derecha. Por lo tanto, la comparación de los datos dentro de cada intervención es problemática, en particular para la EMT prefrontal, en la cual el procedimiento varió mucho más.
5. Resultados
De las siete categorías de resultados predefinidos, seis se analizaron en las intervenciones EMT prefrontal y EMT temporoparietal. No hubo datos disponibles para el análisis en las categorías de resultados hospitalarios y de servicio, la satisfacción con la atención y los resultados económicos. Faltaron los datos sobre la calidad de vida; solamente un estudio informó este resultado para la EMT temporoparietal. Los ensayos futuros deben considerar la posibilidad de incluir mecanismos para obtener estos datos adicionales; sin embargo, los revisores reconocen los riesgos de obtener datos limitados de buena calidad a expensas de la cantidad.
Calidad de la evidencia
Esta revisión incluye 41 estudios con 1473 participantes. Aunque todos los estudios se informaron como aleatorios, la mayoría informó una ocultación incierta de la asignación (Figura 2 y Figura 3). Por lo tanto, existe un riesgo de sesgo de selección debido a la posible falta de buenos métodos para ocultar la asignación. En general, solamente siete artículos informaron la ocultación adecuada de la asignación. La mayoría de los estudios incluidos en la revisión se describió como doble ciego. Sin embargo, solamente siete estudios informaron un procedimiento adecuado para el cegamiento. Este hecho puede influir en el sesgo de realización y en el de desgaste, y es de especial importancia en estos diseños de estudio debido al uso de medidas subjetivas. En muchos estudios los datos no se informaron completamente y fue necesario establecer contacto con los autores para obtener las medias y las desviaciones estándar para las medidas individuales.
En general, la calidad de las pruebas se consideró baja a muy baja. En general, los resultados fueron consistentes, aunque los resultados de los datos de las escalas para el estado mental mostraron cierta heterogeneidad. La calificación de las pruebas en las tablas "Resumen de los hallazgos" se disminuyó principalmente debido a los riesgos de sesgo de los estudios mencionados anteriormente y a la imprecisión de los resultados debido a los amplios intervalos de confianza.
Sesgos potenciales en el proceso de revisión
Se intentó identificar todos los ensayos relevantes con la estrategia de búsqueda. Sin embargo, es posible que no se hayan identificado todos los estudios. También se reconoce que la fecha de búsqueda era antigua en el momento de la publicación y puede haber estudios nuevos disponibles.
La extracción de los datos y las evaluaciones del riesgo de sesgo de los estudios en idioma chino fueron completadas por un solo revisor. Existe la posibilidad de que se pueda haber introducido algún sesgo en los resultados, ya que no fue posible la verificación de estos datos.
Acuerdos y desacuerdos con otros estudios o revisiones
Los estudios de investigación anteriores han encontrado que la EMT puede reducir significativamente los síntomas de la esquizofrenia. Sin embargo, no todos los estudios han replicado posteriormente estos resultados. Este hecho también se reflejó en los resultados de la presente revisión ya que, aunque hubo algunas pruebas para apoyar el efecto beneficioso de la EMT sobre la esquizofrenia, los resultados no fueron consistentes para todas las medidas.
La presente revisión concluyó que hay pruebas limitadas de que la EMT temporoparietal es superior a la EMT simulada para mejorar las alucinaciones auditivas; dos metanálisis de estudios controlados mediante simulación concuerdan con este resultado, y ambos concluyen que hubo un tamaño del efecto grande y significativo en la mejoría de las alucinaciones auditivas (Freitas 2009; Matheson 2010a). La presente revisión encontró pruebas limitadas de que la EMT temporoparietal es superior a la EMT simulada para mejorar los síntomas positivos cuando se miden con la escala PANSS; un metanálisis de estudios controlados mediante simulación concuerdan, y encontró un tamaño del efecto grande y significativo(Freitas 2009).
También se informaron pruebas limitadas de que la EMT prefrontal es superior a la simulada para mejorar los síntomas negativos en dos metanálisis (Dlabac‐de Lange 2010; Freitas 2009), el primero encontró una mejoría estadísticamente significativa en un análisis de PANSS y SANS, y el último encontró un tamaño del efecto pequeño y no significativo en los síntomas negativos. Se alcanzó un acuerdo adicional de que la EMT prefrontal puede ser eficaz para mejorar los síntomas negativos en una comunicación de seguimiento a una revisión(Matheson 2010b).
Study flow diagram.
Risk of bias summary: review authors' judgements about each risk of bias item for each included study.
Risk of bias graph: review authors' judgements about each risk of bias item presented as percentages across all included studies.
Comparison 1 TEMPOROPARIETAL TMS vs SHAM TMS, Outcome 1 Global state: 1. Clinical improvement (CGI).
Comparison 1 TEMPOROPARIETAL TMS vs SHAM TMS, Outcome 2 Global state: 2. Average score for clinical improvement (CGI, high = poor).
Comparison 1 TEMPOROPARIETAL TMS vs SHAM TMS, Outcome 3 Mental state: 1. General: a. Clinical improvement (PANSS > 30% reduction).
Comparison 1 TEMPOROPARIETAL TMS vs SHAM TMS, Outcome 4 Mental state: 1. General: b. Average total score (various scales).
Comparison 1 TEMPOROPARIETAL TMS vs SHAM TMS, Outcome 5 Mental state: 1. General: c. Average general psychopathology score (PANSS general, high = poor).
Comparison 1 TEMPOROPARIETAL TMS vs SHAM TMS, Outcome 6 Mental state: 2. Specific: a. Average depression score (various scales).
Comparison 1 TEMPOROPARIETAL TMS vs SHAM TMS, Outcome 7 Mental state: 2. Specific: b.i. Hallucinations ‐ clinical improvement (various scales).
Comparison 1 TEMPOROPARIETAL TMS vs SHAM TMS, Outcome 8 Mental state: 2. Specific: b.ii. Average hallucinations score (various scales).
Comparison 1 TEMPOROPARIETAL TMS vs SHAM TMS, Outcome 9 Mental state: 2. Specific: c. Average negative symptom score (various scales).
Comparison 1 TEMPOROPARIETAL TMS vs SHAM TMS, Outcome 10 Mental state: 2. Specific: d.i. Positive symptoms ‐ clinical improvement (PANSS > 30% reduction).
Comparison 1 TEMPOROPARIETAL TMS vs SHAM TMS, Outcome 11 Mental state: 2. Specific: d.ii. Average positive symptom score (various scales).
Comparison 1 TEMPOROPARIETAL TMS vs SHAM TMS, Outcome 12 Adverse effects: 1. General: a. Serious.
Comparison 1 TEMPOROPARIETAL TMS vs SHAM TMS, Outcome 13 Adverse effects: 1. General: b. Leaving the study early.
Comparison 1 TEMPOROPARIETAL TMS vs SHAM TMS, Outcome 14 Adverse effects: 2. Specific.
Comparison 1 TEMPOROPARIETAL TMS vs SHAM TMS, Outcome 15 Quality of life: Average score (Q‐LES‐Q, low = poor).
Comparison 2 TEMPOROPARIETAL TMS vs STANDARD TREATMENT, Outcome 1 Global state: Clinical improvement (CGI ≤ 2).
Comparison 2 TEMPOROPARIETAL TMS vs STANDARD TREATMENT, Outcome 2 Adverse effects: Leaving the study early.
Comparison 3 PREFRONTAL TMS vs SHAM TMS, Outcome 1 Global state: Average score (various scales).
Comparison 3 PREFRONTAL TMS vs SHAM TMS, Outcome 2 Mental state: 1. General: a. Clinical improvement (> 20% decrease in total PANSS score).
Comparison 3 PREFRONTAL TMS vs SHAM TMS, Outcome 3 Mental state: 1. General: b. Average total score (various scales).
Comparison 3 PREFRONTAL TMS vs SHAM TMS, Outcome 4 Mental state: 1. General: c. Average general psychopathology score (PANSS, high = poor).
Comparison 3 PREFRONTAL TMS vs SHAM TMS, Outcome 5 Mental state: 2. Specific: a. Average depression score (various scales).
Comparison 3 PREFRONTAL TMS vs SHAM TMS, Outcome 6 Mental state: 2. Specific: b. Average hallucinations score (PANSS, high = poor).
Comparison 3 PREFRONTAL TMS vs SHAM TMS, Outcome 7 Mental state: 2. Specific: c. i. Negative symptoms ‐ clinical improvement (> 20% decrease in PANSS negative).
Comparison 3 PREFRONTAL TMS vs SHAM TMS, Outcome 8 Mental state: 2. Specific: c. ii. Average negative symptom score (various scales).
Comparison 3 PREFRONTAL TMS vs SHAM TMS, Outcome 9 Mental state: 2. Specific: d. Average positive symptom score (various scales).
Comparison 3 PREFRONTAL TMS vs SHAM TMS, Outcome 10 Mental state: 2. Specific: e. Average psychotism score (SCL‐90 PSY, high = poor).
Comparison 3 PREFRONTAL TMS vs SHAM TMS, Outcome 11 Adverse effects: 1. General: a. Adverse events (UKU).
Comparison 3 PREFRONTAL TMS vs SHAM TMS, Outcome 12 Adverse effects: 1. General: b. Leaving the study early.
Comparison 3 PREFRONTAL TMS vs SHAM TMS, Outcome 13 Adverse effects: 2. Specific: a. Various.
Comparison 3 PREFRONTAL TMS vs SHAM TMS, Outcome 14 Adverse effects: 2. Specific: b. Average score (CSSES, high = poor).
Comparison 4 PREFRONTAL THETA BURST STIMULATION TMS vs SHAM TMS, Outcome 1 Global state: Clinical improvement.
Comparison 4 PREFRONTAL THETA BURST STIMULATION TMS vs SHAM TMS, Outcome 2 Mental state: 1. General: a. Average overall mental state score (PANSS total, high = poor).
Comparison 4 PREFRONTAL THETA BURST STIMULATION TMS vs SHAM TMS, Outcome 3 Mental state: 1. General: b. Average general psychopathology score (PANSS, high = poor).
Comparison 4 PREFRONTAL THETA BURST STIMULATION TMS vs SHAM TMS, Outcome 4 Mental state: 2. Specific: a. Average negative symptom score (various scales).
Comparison 4 PREFRONTAL THETA BURST STIMULATION TMS vs SHAM TMS, Outcome 5 Mental state: 2. Specific: b. Average positive symptom score (PANSS, high = poor).
Comparison 4 PREFRONTAL THETA BURST STIMULATION TMS vs SHAM TMS, Outcome 6 Cognitive state: Average score (various measures).
Comparison 4 PREFRONTAL THETA BURST STIMULATION TMS vs SHAM TMS, Outcome 7 Adverse effects: 1. Leaving the study early.
Comparison 4 PREFRONTAL THETA BURST STIMULATION TMS vs SHAM TMS, Outcome 8 Adverse effects: 2. Specific.
Comparison 5 SENSITIVITY ANALYSIS: PREFRONTAL THETA BURST STIMULATION TMS vs SHAM TMS, Outcome 1 Global state: Clinical improvement.
| TEMPOROPARIETAL TMS compared to SHAM TMS for schizophrenia | ||||||
| Patient or population: people with schizophrenia | ||||||
| Outcomes | Illustrative comparative risks* (95% CI) | Relative effect | No of Participants | Quality of the evidence | Comments | |
| Assumed risk | Corresponding risk | |||||
| SHAM TMS | TEMPOROPARIETAL TMS | |||||
| Clinical improvement in global state | 0 per 1000 | 0 per 1000 | RR 7 | 46 | ⊕⊝⊝⊝ | |
| Global state score | The mean scores for the sham TMS group ranged from 2.4 to 5.1 | The mean global state score in the intervention groups was | 224 | ⊕⊕⊝⊝ | ||
| Mental state | The mean mental state scores in the sham TMS group ranged from 63.92 to 85.8 | The mean mental state in the intervention groups was | 127 | ⊕⊕⊝⊝ | ||
| Cognitive state | See comment | See comment | Not estimable | 82 | ⊕⊕⊝⊝ | Cognitive state was reported in 3 studies using 39 different measures. Results were equivocal for all measures |
| Adverse effects: general or specific | See comment | See comment | Not estimable | 442 | ⊕⊕⊝⊝ | There were more headaches and jaw and facial contraction in the TMS group. Results for other adverse events ‐ concentration problems, earache, lightheadedness, mild amnesia, restless legs, somatic discomfort, tingling sensation in the arm, worsening hallucinations ‐ were equivocal |
| Adverse effects: Leaving the study early | 152 per 1000 | 118 per 1000 | RR 0.78 | 320 | ⊕⊝⊝⊝ | |
| Satisfaction with care ‐ not reported | See comment | See comment | Not estimable | ‐ | See comment | No studies reported on this outcome |
| *The basis for the assumed risk is the median control group risk across studies for pooled data and the control group risk for single studies. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). | ||||||
| GRADE Working Group grades of evidence | ||||||
| 1Risk of bias: serious ‐ this study had an unclear risk of bias for randomisation, allocation concealment, blinding and incomplete outcome data. Downgraded one level. | ||||||
| TEMPOROPARIETAL TMS compared to STANDARD TREATMENT for schizophrenia | ||||||
| Patient or population: people with schizophrenia | ||||||
| Outcomes | Illustrative comparative risks* (95% CI) | Relative effect | No of Participants | Quality of the evidence | Comments | |
| Assumed risk | Corresponding risk | |||||
| STANDARD TREATMENT | TEMPOROPARIETAL TMS | |||||
| Clinical improvement in global state | 620 per 1000 | 738 per 1000 | RR 1.19 | 100 | ⊕⊕⊝⊝ | |
| Global state score ‐ not reported | See comment | See comment | Not estimable | ‐ | See comment | No studies reported data for this outcome |
| Mental state ‐ not reported | See comment | See comment | Not estimable | ‐ | See comment | No studies reported data for this outcome |
| Cognitive state ‐ not reported | See comment | See comment | Not estimable | ‐ | See comment | No studies reported data for this outcome |
| Adverse effects: general or specific ‐ not reported | See comment | See comment | Not estimable | ‐ | See comment | No studies reported data for this outcome |
| Adverse effects: Leaving the study early | 86 per 1000 | 28 per 1000 | RR 0.33 | 140 | ⊕⊝⊝⊝ | |
| Satisfaction with care ‐ not reported | See comment | See comment | Not estimable | ‐ | See comment | No studies reported data for this outcome |
| *The basis for the assumed risk is the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). | ||||||
| GRADE Working Group grades of evidence | ||||||
| 1Participants received first and second generation antipsychotics in both treatment groups | ||||||
| PREFRONTAL TMS compared to SHAM TMS for schizophrenia | ||||||
| Patient or population: people with schizophrenia | ||||||
| Outcomes | Illustrative comparative risks* (95% CI) | Relative effect | No of Participants | Quality of the evidence | Comments | |
| Assumed risk | Corresponding risk | |||||
| SHAM TMS | PREFRONTAL TMS | |||||
| Clinical improvement in global state ‐ not reported | See comment | See comment | Not estimable | ‐ | See comment | No studies reported data for this outcome |
| Global state score | See comment | See comment | Not estimable | 85 | See comment | 3 small studies measured global state on the CGI, CGI‐S, GAF and SCL‐90, none of which showed a significant treatment effect |
| Mental state | See comment | See comment | Not estimable | 188 | ⊕⊝⊝⊝ | There was very high heterogeneity for this outcome, so we did not pool the data |
| Cognitive state | See comment | See comment | Not estimable | 138 | ⊕⊕⊝⊝ | Cognitive state was reported in 4 studies using 19 different measures. Results were equivocal for all measures apart from 4 |
| Adverse effects: general or specific | See comment | See comment | Not estimable | 199 | ⊕⊕⊝⊝ | There were more headaches and TMS‐related site discomfort or pain in the TMS group. Results for other adverse effects ‐ cognitive difficulties, facial twitching, worsening of pre‐existing akathisia and OCD ‐ were equivocal |
| Adverse effects: Leaving the study early | 106 per 1000 | 126 per 1000 | RR 1.19 | 174 | ⊕⊕⊝⊝ | |
| Satisfaction with care ‐ not reported | See comment | See comment | Not estimable | ‐ | See comment | No studies reported data for this outcome |
| *The basis for the assumed risk (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). | ||||||
| GRADE Working Group grades of evidence | ||||||
| 1Risk of bias: serious ‐ two studies had an unclear risk of bias for randomisation and five for allocation concealment. Five studies had an unclear risk for blinding of participants and two for blinding of outcome assessors and incomplete outcome data. | ||||||
| PREFRONTAL TBS TMS compared to SHAM TMS for schizophrenia | ||||||
| Patient or population: people with schizophrenia | ||||||
| Outcomes | Illustrative comparative risks* (95% CI) | Relative effect | No of Participants | Quality of the evidence | Comments | |
| Assumed risk | Corresponding risk | |||||
| SHAM TMS | PREFRONTAL TBS TMS | |||||
| Clinical improvement in global state | 0 per 1000 | 0 per 1000 | RR 4.06 | 27 | ⊕⊝⊝⊝ | |
| Global state score ‐ not reported | See comment | See comment | Not estimable | ‐ | See comment | No studies reported data for this outcome |
| Mental state | The mean scores for the sham TMS group ranged from 67.6 to 68.4 | The mean mental state in the intervention groups was | 108 | ⊕⊕⊝⊝ | ||
| Cognitive state | See comment | See comment | Not estimable | 39 | ⊕⊕⊝⊝ | This was measured on 2 scale, both showed equivocal results |
| Adverse effects: general or specific | See comment | See comment | Not estimable | 27 | ⊕⊕⊝⊝ | No differences in headaches and sleep disorders |
| Adverse effects: Leaving the study early | 139 per 1000 | 50 per 1000 | RR 0.36 | 76 | ⊕⊕⊝⊝ | |
| Satisfaction with care ‐ not reported | See comment | See comment | Not estimable | ‐ | See comment | No studies reported data for this outcome |
| *The basis for the assumed risk (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). | ||||||
| GRADE Working Group grades of evidence | ||||||
| 1Risk of bias: serious ‐ there was an unclear risk of bias for allocation concealment and blinding | ||||||
| Contact details |
| Mr John H Starzewski |
| Study | Outcome | TMS Mean | TMS SD | TMS N | Sham TMS Mean | Sham TMS SD | Sham TMS N |
| Mental state: specific – BPRS depressive factor (high = poor) | 2.25 | 2.18 | 8 | 3.56 | 3.24 | 9 | |
| Mental state: specific – BPRS excitement factor (high = poor) | 1.25 | 1.28 | 8 | 3.89 | 4.79 | 9 | |
| Mental state: Specific ‐ positive symptoms (SAPS, high = poor) | 51.2 | 13.5 | 5 | 47.8 | 25.2 | 5 | |
| Mental state: Specific ‐ hallucinations (AHRS, high = poor) | 14.6 | 12.1 | 5 | 20.8 | 3.4 | 5 | |
| Mental state: Specific ‐ positive symptoms (SAPS, high = poor) | 26 | 20 | 10 | 37 | 16 | 10 | |
| Mental state: Specific ‐ negative symptoms (SANS, high = poor) | 32 | 27 | 10 | 39 | 23 | 10 | |
| Cognitive state: CPT false items | 1.94 | 2.04 | 18 | 1.41 | 2.12 | 17 | |
| Cognitive state: CPT missed items | 6.28 | 4.5 | 18 | 7.59 | 6.68 | 17 | |
| AHRS ‐ Auditory Hallucination Rating Scale | |||||||
| Outcome | Change / endpoint data | Study | TMS | Sham TMS | Mean difference [95% CI] | ||||
| Mean | SD | N | Mean | SD | N | ||||
| Animal naming | Change | ‐0.77 | 4.41 | 26 | 0.9 | 4.17 | 21 | ‐1.67 [‐4.13 to 0.79] | |
| CPT reaction time (ms) | Endpoint | 926.22 | 126.2 | 18 | 959 | 109.35 | 17 | ‐32.78 [‐110.89 to 45.33] | |
| Controlled oral word association | Change | 2.57 | 7.07 | 26 | 2.53 | 0.91 | 21 | 0.04 [‐2.71 to 2.79] | |
| CVLT 1 score | Change | 0.88 | 6.61 | 26 | ‐0.19 | 1.69 | 21 | 1.07 [‐1.57 to 3.71] | |
| CVLT B score | Change | 0.15 | 1.76 | 26 | 0.48 | 3.1 | 21 | ‐0.33 [‐1.82 to 1.16] | |
| CVLT Long‐delay free recall | Change | ‐1.69 | 2.28 | 26 | ‐1.48 | 3.1 | 21 | ‐0.21 [‐1.80 to 1.38] | |
| CVLT Recognition discriminative ability | Change | ‐0.007 | 0.08 | 26 | ‐0.014 | 0.088 | 21 | 0.01 [‐0.04 to 0.06] | |
| CVLT Short‐delay free recall | Change | ‐0.92 | 2.15 | 26 | ‐0.71 | 3 | 21 | ‐0.21 [‐1.74 to 1.32] | |
| CVLT1‐5 Total score | Change | ‐3.42 | 7.08 | 26 | ‐3.14 | 8.21 | 21 | ‐0.28 [‐4.72 to 4.16] | |
| Digit recall (distraction) | Change | 0.61 | 3.93 | 26 | ‐0.9 | 4.77 | 21 | 1.51 [‐1.03 to 4.05] | |
| Digit recall (non‐distraction) | Change | ‐0.12 | 4.66 | 26 | 1.19 | 4.08 | 21 | ‐1.31 [‐3.81 to 1.19] | |
| Digit symbol | Change | 3.15 | 7.76 | 26 | 2.95 | 7.72 | 21 | 0.20 [‐4.25 to 4.65] | |
| Grooved pegboard, dominant | Change | 4.65 | 15.1 | 26 | 5.57 | 48.8 | 21 | ‐0.92 [‐22.58 to 20.74] | |
| Grooved pegboard, nondominant | Change | 6.46 | 15.5 | 26 | 12 | 31.5 | 21 | ‐5.54 [‐20.27 to 9.19] | |
| Temporal orientation | Change | ‐0.154 | 1.82 | 26 | ‐0.35 | 2.39 | 21 | 0.20 [‐1.04 to 1.43] | |
| Trail Making A | Change | 2.58 | 12.6 | 26 | ‐0.42 | 8.23 | 21 | 3.00 [‐2.99 to 8.99] | |
| Trail Making B | Change | 19.5 | 48.3 | 26 | 25.3 | 50.4 | 21 | ‐5.80 [‐34.25 to 22.65] | |
| WCST completed categories | Endpoint | 2.17 | 2.23 | 18 | 2.82 | 2.32 | 17 | ‐0.65 [‐2.16 to 0.86] | |
| WCST completed categories | Endpoint | 5.3 | 1 | 11 | 4.5 | 1.4 | 10 | 0.80 [‐0.25 to 1.85] | |
| WCST conceptualisation level | Endpoint | 61 | 24.13 | 18 | 64.12 | 24.93 | 17 | ‐3.12 [‐19.39 to 13.15] | |
| WCST CR | Endpoint | 54 | 9 | 11 | 49 | 11 | 10 | 5.00 [‐3.65 to 13.65] | |
| WCST FM | Endpoint | 13.6 | 7.3 | 11 | 9.8 | 11.2 | 10 | 3.80 [‐4.37 to 11.97] | |
| WCST NPE | Endpoint | 17.2 | 7.6 | 11 | 20.9 | 5.6 | 10 | ‐3.70 [‐9.38 to 1.98] | |
| WCST PCLR | Endpoint | 60.5 | 19.2 | 11 | 45.9 | 18.6 | 10 | 14.60 [‐1.58 to 30.78] | |
| WCST PE | Endpoint | 45.3 | 23.6 | 11 | 50.8 | 22.3 | 10 | ‐5.50 [‐25.14 to 14.14] | |
| WCST PNPE | Endpoint | 31.3 | 13.8 | 11 | 32.4 | 14.5 | 10 | ‐1.10 [‐13.24 to 11.04] | |
| WCST PPE | Endpoint | 68.7 | 13.8 | 11 | 67.6 | 14.6 | 10 | 1.10 [‐11.08 to 13.28] | |
| WCST PR | Endpoint | 41 | 13.3 | 11 | 34.2 | 15 | 10 | 6.80 [‐5.37 to 18.97] | |
| WCST Ra | Endpoint | 122.67 | 15.18 | 18 | 126.06 | 5.02 | 17 | ‐3.39 [‐10.80 to 4.02] | |
| WCST Re | Endpoint | 56.11 | 22.99 | 18 | 53.88 | 16.14 | 17 | 2.23 [‐10.87 to 15.33] | |
| WCST Rp | Endpoint | 45.72 | 20.18 | 18 | 60.12 | 19.23 | 17 | ‐14.40 [‐27.46 to ‐1.34] | |
| WCST TA | Endpoint | 117 | 18 | 11 | 121 | 10 | 10 | ‐4.00 [‐16.31 to 8.31] | |
| WCST TCFC | Endpoint | 21.7 | 14 | 11 | 29.0 | 13.4 | 10 | ‐7.30 [‐19.02 to 4.42] | |
| WCST TE | Endpoint | 63 | 24 | 11 | 72 | 20 | 10 | ‐9.00 [‐27.84 to 9.84] | |
| WCST time (sec) | Endpoint | 405 | 174 | 11 | 411 | 177 | 10 | ‐6.00 [‐156.36 to 144.36] | |
| WCST selective error rate (%) | Endpoint | 51.6 | 15.3 | 11 | 58.4 | 12.3 | 10 | ‐6.80 [‐18.63 to 5.03] | |
| WCST correct thinking time (sec) | Endpoint | 172 | 67 | 11 | 160 | 96 | 10 | 12.00 [‐59.47 to 83.47] | |
| WCST error thinking time (sec) | Endpoint | 233 | 128 | 11 | 251 | 100 | 10 | ‐18.00 [‐115.79 to 79.79] | |
| WRAT‐R | Change | 0.19 | 2.54 | 26 | 0.33 | 2.81 | 21 | 6.80 [‐5.37 to 18.97] | |
| CPT ‐ Continuous performance test | |||||||||
| Study | Outcome | TMS Mean | TMS SD | TMS N | Sham TMS Mean | Sham TMS SD | Sham TMS N |
| Mental state: Specific ‐ hallucinations (AHRS, high = poor) | 6.7 | 8.64 | 20 | 27.95 | 7.51 | 20 | |
| AHRS ‐ Auditory Hallucination Rating Scale | |||||||
| Study | Outcome | TMS Mean | TMS SD | TMS N | Sham TMS Mean | Sham TMS SD | Sham TMS N |
| Mental state: Specific ‐ negative symptoms (PANSS, high = poor) | 26.15 | 13.45 | 13 | 31.42 | 13.19 | 12 | |
| Mental state: Specific – depressive symptoms (CDS, high = poor ) | 2.38 | 2.06 | 13 | 1.67 | 1.92 | 12 | |
| Mental state: Specific ‐ positive symptoms (PANSS, high = poor) (LOCF) | 10.8 | 7.0 | 10 | 7.3 | 2.9 | 10 | |
| Mental state: Depression (CDRS, high = poor) (LOCF) | 7.2 | 5.9 | 10 | 3.5 | 3.8 | 10 | |
| CDRS ‐ Calgary depression rating scale | |||||||
| Outcome | Change / endpoint | Study | TMS | Sham TMS | Mean difference [95% CI] | ||||
| Mean | SD | N | Mean | SD | N | ||||
| AVLT (low = poor) | Endpoint | 45.6 | 6.8 | 8 | 44.9 | 8 | 8 | 0.70 [‐6.58 to 7.98] | |
| COWAT (within 24 hours of TMS) | Endpoint | 11.6 | 5.3 | 8 | 10.9 | 5.0 | 9 | 0.70 [‐4.22 to 5.62] | |
| COWAT (2 weeks after TMS) | Endpoint | 14.2 | 5.7 | 8 | 9.1 | 2.7 | 9 | 5.10 [0.77 to 9.43] | |
| Digit span test | Endpoint | 10.5 | 3.5763 | 38 | 9.5 | 4 | 17 | 1.00 [‐15.70 to 17.70] | |
| Grooved pegboard (seconds to completion) (within 24 hours of TMS) | Endpoint | 117.1 | 32.0 | 8 | 108.6 | 41.2 | 9 | 8.50 [‐26.37 to 43.37] | |
| Grooved pegboard (seconds to completion) (2 weeks after TMS) | Endpoint | 109 | 29.5 | 8 | 98.5 | 16 | 9 | 10.50 [‐12.46 to 33.46] | |
| HVLT‐ delayed recall (within 24 hours of TMS) | Endpoint | 4.4 | 2.3 | 8 | 4.4 | 1.1 | 9 | 0.00 [‐1.75 to 1.75] | |
| HVLT‐ delayed recall (2 weeks after TMS) | Endpoint | 5.4 | 2.7 | 8 | 3.3 | 1.0 | 9 | 2.10 [0.12 to 4.08] | |
| HVLT‐ immediate recall (within 24 hours of TMS) | Endpoint | 6.3 | 2.0 | 8 | 5.6 | 1.1 | 9 | 0.70 [‐0.86 to 2.26] | |
| HVLT‐ immediate recall (2 weeks after TMS) | Endpoint | 7.4 | 2.8 | 8 | 5 | 0.8 | 9 | 2.40 [0.39 to 4.41] | |
| Stroop test (within 24 hours of TMS) | Endpoint | 77.4 | 20.3 | 8 | 51.4 | 14.9 | 9 | 26.00 [8.89 to 43.11] | |
| Stroop test (2 weeks after TMS) | Endpoint | 88.2 | 12.3 | 8 | 60.8 | 6.4 | 9 | 27.40 [17.91 to 36.89] | |
| Trail making test A | Change | 0.64 | 15.08 | 14 | ‐11.92 | 29.27 | 12 | 12.56 [‐5.79 to 30.91] | |
| Trail making test B | Change | ‐0.54 | 28.41 | 13 | ‐5.64 | 20.31 | 11 | 5.10 [‐14.46 to 24.66] | |
| Verbal fluency test (high = poor) | Endpoint | 24.2 | 9.2542 | 38 | 24.5 | 6.6 | 17 | ‐0.30 [‐4.60 to 4.00] | |
| WCST categories | Change | 1.58 | 22.2 | 12 | ‐0.27 | 1.95 | 11 | 1.85 [‐10.76 to 14.46] | |
| WCST categories for participants with WCST categories pre < median (= 4) | Change | 3.33 | 2.58 | 6 | 0.4 | 2.07 | 5 | 2.93 [0.18 to 5.68] | |
| WCST perseverative answers | Change | ‐9 | 11.65 | 12 | ‐19.18 | 27.76 | 11 | 10.18 [‐7.50 to 27.86] | |
| WCST perseverative mistakes | Change | ‐8.17 | 9.81 | 12 | ‐11.27 | 17.51 | 11 | 3.10 [‐8.64 to 14.84] | |
| AVTL ‐ auditory verbal learning test | |||||||||
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 1 Global state: 1. Clinical improvement (CGI) Show forest plot | 1 | 46 | Risk Ratio (M‐H, Fixed, 95% CI) | 7.0 [0.38, 128.33] |
| 2 Global state: 2. Average score for clinical improvement (CGI, high = poor) Show forest plot | 7 | 224 | Mean Difference (IV, Fixed, 95% CI) | ‐0.50 [‐0.76, ‐0.23] |
| 3 Mental state: 1. General: a. Clinical improvement (PANSS > 30% reduction) Show forest plot | 1 | 51 | Risk Ratio (M‐H, Fixed, 95% CI) | 1.0 [0.10, 10.27] |
| 4 Mental state: 1. General: b. Average total score (various scales) Show forest plot | 6 | Mean Difference (IV, Fixed, 95% CI) | Subtotals only | |
| 4.1 BPRS (high = poor) | 1 | 17 | Mean Difference (IV, Fixed, 95% CI) | ‐5.68 [‐12.98, 1.62] |
| 4.2 PANSS total (high = poor) | 5 | 127 | Mean Difference (IV, Fixed, 95% CI) | ‐6.09 [‐10.95, ‐1.22] |
| 5 Mental state: 1. General: c. Average general psychopathology score (PANSS general, high = poor) Show forest plot | 4 | 87 | Mean Difference (IV, Fixed, 95% CI) | ‐2.34 [‐5.26, 0.59] |
| 6 Mental state: 2. Specific: a. Average depression score (various scales) Show forest plot | 1 | Mean Difference (IV, Fixed, 95% CI) | Subtotals only | |
| 6.1 HAMD (high = poor) | 1 | 25 | Mean Difference (IV, Fixed, 95% CI) | ‐3.92 [‐7.84, ‐0.00] |
| 6.2 SDS (high = poor) | 1 | 25 | Mean Difference (IV, Fixed, 95% CI) | ‐5.59 [‐11.57, 0.39] |
| 7 Mental state: 2. Specific: b.i. Hallucinations ‐ clinical improvement (various scales) Show forest plot | 5 | Risk Ratio (M‐H, Fixed, 95% CI) | Subtotals only | |
| 7.1 AHRS >30% decrease in symptoms | 3 | 120 | Risk Ratio (M‐H, Fixed, 95% CI) | 2.99 [1.12, 7.98] |
| 7.2 HCS score ≤5 | 3 | 133 | Risk Ratio (M‐H, Fixed, 95% CI) | 2.26 [1.18, 4.35] |
| 7.3 PANSS hallucination item improvement ≥1 point | 1 | 36 | Risk Ratio (M‐H, Fixed, 95% CI) | 1.33 [0.43, 4.13] |
| 7.4 PSYRATS > 30% reduction | 1 | 51 | Risk Ratio (M‐H, Fixed, 95% CI) | 3.6 [0.20, 65.96] |
| 8 Mental state: 2. Specific: b.ii. Average hallucinations score (various scales) Show forest plot | 13 | Mean Difference (IV, Random, 95% CI) | Subtotals only | |
| 8.1 AHRS (high = poor) | 9 | 327 | Mean Difference (IV, Random, 95% CI) | ‐2.11 [‐4.38, 0.16] |
| 8.2 AVH‐related items PSYRATS (high = poor) | 2 | 624 | Mean Difference (IV, Random, 95% CI) | ‐0.51 [‐3.38, 2.36] |
| 8.3 HCS (high = poor) | 3 | 162 | Mean Difference (IV, Random, 95% CI) | ‐1.64 [‐2.80, ‐0.48] |
| 8.4 PANSS hallucination item (high = poor) | 4 | 125 | Mean Difference (IV, Random, 95% CI) | ‐1.01 [‐1.97, ‐0.04] |
| 9 Mental state: 2. Specific: c. Average negative symptom score (various scales) Show forest plot | 8 | Mean Difference (IV, Fixed, 95% CI) | Subtotals only | |
| 9.1 BPRS (high = poor) | 1 | 17 | Mean Difference (IV, Fixed, 95% CI) | ‐3.06 [‐7.15, 1.03] |
| 9.2 PANSS (high = poor) | 7 | 162 | Mean Difference (IV, Fixed, 95% CI) | ‐0.31 [‐1.87, 1.25] |
| 9.3 SANS (high = poor) | 1 | 25 | Mean Difference (IV, Fixed, 95% CI) | ‐23.58 [‐37.06, ‐10.10] |
| 10 Mental state: 2. Specific: d.i. Positive symptoms ‐ clinical improvement (PANSS > 30% reduction) Show forest plot | 1 | 51 | Risk Ratio (M‐H, Fixed, 95% CI) | 1.0 [0.10, 10.27] |
| 11 Mental state: 2. Specific: d.ii. Average positive symptom score (various scales) Show forest plot | 13 | Mean Difference (IV, Fixed, 95% CI) | Subtotals only | |
| 11.1 BPRS (high = poor) | 1 | 17 | Mean Difference (IV, Fixed, 95% CI) | 0.53 [‐2.78, 3.84] |
| 11.2 PANSS (high = poor) | 11 | 333 | Mean Difference (IV, Fixed, 95% CI) | ‐2.14 [‐3.15, ‐1.14] |
| 11.3 SAPS (high = poor) | 2 | 49 | Mean Difference (IV, Fixed, 95% CI) | ‐3.22 [‐7.86, 1.42] |
| 12 Adverse effects: 1. General: a. Serious Show forest plot | 2 | 130 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] |
| 13 Adverse effects: 1. General: b. Leaving the study early Show forest plot | 8 | 320 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.78 [0.46, 1.32] |
| 14 Adverse effects: 2. Specific Show forest plot | 11 | Risk Ratio (M‐H, Fixed, 95% CI) | Subtotals only | |
| 14.1 cardiovascular ‐ lightheaded/Dizziness | 3 | 158 | Risk Ratio (M‐H, Fixed, 95% CI) | 1.60 [0.45, 5.75] |
| 14.2 central nervous system ‐ tinnitus | 1 | 83 | Risk Ratio (M‐H, Fixed, 95% CI) | 3.63 [0.19, 67.82] |
| 14.3 cognitive ‐ concentration problems | 2 | 122 | Risk Ratio (M‐H, Fixed, 95% CI) | 1.59 [0.26, 9.73] |
| 14.4 cognitive ‐ mild memory impairment/amnesia | 2 | 89 | Risk Ratio (M‐H, Fixed, 95% CI) | 2.90 [0.35, 24.18] |
| 14.5 movement disorder ‐ jaw and facial contraction | 2 | 70 | Risk Ratio (M‐H, Fixed, 95% CI) | 8.32 [1.13, 61.17] |
| 14.6 movement disorder ‐ restless legs | 1 | 36 | Risk Ratio (M‐H, Fixed, 95% CI) | 1.56 [0.07, 35.67] |
| 14.7 psychiatric ‐ worsening hallucinations/audible Thoughts | 1 | 83 | Risk Ratio (M‐H, Fixed, 95% CI) | 2.55 [0.31, 20.75] |
| 14.8 others ‐ earache | 1 | 36 | Risk Ratio (M‐H, Fixed, 95% CI) | 1.56 [0.07, 35.67] |
| 14.9 others ‐ headache | 10 | 392 | Risk Ratio (M‐H, Fixed, 95% CI) | 2.65 [1.56, 4.50] |
| 14.10 others ‐ somatic discomfort | 1 | 83 | Risk Ratio (M‐H, Fixed, 95% CI) | 1.78 [0.65, 4.91] |
| 14.11 others ‐ tingling sensation in the arm | 1 | 36 | Risk Ratio (M‐H, Fixed, 95% CI) | 1.56 [0.07, 35.67] |
| 15 Quality of life: Average score (Q‐LES‐Q, low = poor) Show forest plot | 1 | 20 | Mean Difference (IV, Fixed, 95% CI) | ‐1.0 [‐14.26, 12.26] |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 1 Global state: Clinical improvement (CGI ≤ 2) Show forest plot | 1 | 100 | Risk Ratio (M‐H, Fixed, 95% CI) | 1.19 [0.91, 1.57] |
| 2 Adverse effects: Leaving the study early Show forest plot | 2 | 140 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.33 [0.08, 1.46] |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 1 Global state: Average score (various scales) Show forest plot | 3 | Mean Difference (IV, Fixed, 95% CI) | Subtotals only | |
| 1.1 CGI (high = poor) | 1 | 31 | Mean Difference (IV, Fixed, 95% CI) | 0.60 [‐0.15, 1.35] |
| 1.2 CGI‐S (high = poor) | 1 | 32 | Mean Difference (IV, Fixed, 95% CI) | ‐0.09 [‐0.63, 0.45] |
| 1.3 GAF (low = poor) | 1 | 32 | Mean Difference (IV, Fixed, 95% CI) | 3.43 [‐5.22, 12.08] |
| 1.4 SCL‐90 GSI (high = poor) | 1 | 22 | Mean Difference (IV, Fixed, 95% CI) | ‐0.05 [‐0.66, 0.56] |
| 2 Mental state: 1. General: a. Clinical improvement (> 20% decrease in total PANSS score) Show forest plot | 1 | 22 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.14 [0.02, 0.98] |
| 3 Mental state: 1. General: b. Average total score (various scales) Show forest plot | 7 | Mean Difference (IV, Random, 95% CI) | Totals not selected | |
| 3.1 BPRS (high = poor) | 1 | Mean Difference (IV, Random, 95% CI) | 0.0 [0.0, 0.0] | |
| 3.2 PANSS (high = poor) | 6 | Mean Difference (IV, Random, 95% CI) | 0.0 [0.0, 0.0] | |
| 4 Mental state: 1. General: c. Average general psychopathology score (PANSS, high = poor) Show forest plot | 6 | Mean Difference (IV, Random, 95% CI) | Subtotals only | |
| 5 Mental state: 2. Specific: a. Average depression score (various scales) Show forest plot | 4 | Mean Difference (IV, Fixed, 95% CI) | Subtotals only | |
| 5.1 HAMD‐17 (high = poor) | 1 | 43 | Mean Difference (IV, Fixed, 95% CI) | ‐2.40 [‐3.88, ‐0.92] |
| 5.2 HDRS (high = poor) | 1 | 31 | Mean Difference (IV, Fixed, 95% CI) | 1.70 [‐0.95, 4.35] |
| 5.3 MADRS (high = poor) | 1 | 22 | Mean Difference (IV, Fixed, 95% CI) | ‐4.36 [‐7.05, ‐1.67] |
| 5.4 SCL‐90 DEP (high = poor) | 1 | 22 | Mean Difference (IV, Fixed, 95% CI) | 0.01 [‐0.61, 0.63] |
| 6 Mental state: 2. Specific: b. Average hallucinations score (PANSS, high = poor) Show forest plot | 1 | 25 | Mean Difference (IV, Fixed, 95% CI) | ‐0.68 [‐1.68, 0.32] |
| 7 Mental state: 2. Specific: c. i. Negative symptoms ‐ clinical improvement (> 20% decrease in PANSS negative) Show forest plot | 1 | 16 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.25 [0.04, 1.77] |
| 8 Mental state: 2. Specific: c. ii. Average negative symptom score (various scales) Show forest plot | 13 | Mean Difference (IV, Random, 95% CI) | Subtotals only | |
| 8.1 PANSS (high = poor) | 12 | 341 | Mean Difference (IV, Random, 95% CI) | ‐1.59 [‐4.68, 1.50] |
| 8.2 SANS (high = poor) | 3 | 71 | Mean Difference (IV, Random, 95% CI) | ‐12.68 [‐18.60, ‐6.77] |
| 9 Mental state: 2. Specific: d. Average positive symptom score (various scales) Show forest plot | 10 | Mean Difference (IV, Fixed, 95% CI) | Subtotals only | |
| 9.1 PANSS (high = poor) | 10 | 279 | Mean Difference (IV, Fixed, 95% CI) | ‐0.33 [‐0.99, 0.33] |
| 9.2 SAPS (high = poor) | 1 | 22 | Mean Difference (IV, Fixed, 95% CI) | ‐0.27 [‐2.61, 2.07] |
| 10 Mental state: 2. Specific: e. Average psychotism score (SCL‐90 PSY, high = poor) Show forest plot | 1 | 22 | Mean Difference (IV, Fixed, 95% CI) | ‐0.01 [‐0.48, 0.46] |
| 11 Adverse effects: 1. General: a. Adverse events (UKU) Show forest plot | 1 | 35 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] |
| 12 Adverse effects: 1. General: b. Leaving the study early Show forest plot | 8 | 174 | Risk Ratio (M‐H, Fixed, 95% CI) | 1.19 [0.56, 2.50] |
| 13 Adverse effects: 2. Specific: a. Various Show forest plot | 6 | Risk Ratio (M‐H, Fixed, 95% CI) | Subtotals only | |
| 13.1 cognition ‐ cognitive difficulties | 1 | 31 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] |
| 13.2 movement disorder ‐ facial twitching | 1 | 31 | Risk Ratio (M‐H, Fixed, 95% CI) | 6.59 [0.37, 117.77] |
| 13.3 movement disorder ‐ worsening of pre‐existing akathesia | 1 | 31 | Risk Ratio (M‐H, Fixed, 95% CI) | 4.71 [0.24, 90.69] |
| 13.4 psychiatric ‐ worsening of pre‐existing OCD | 1 | 31 | Risk Ratio (M‐H, Fixed, 95% CI) | 4.71 [0.24, 90.69] |
| 13.5 other ‐ headache | 6 | 164 | Risk Ratio (M‐H, Fixed, 95% CI) | 2.77 [1.22, 6.26] |
| 13.6 other ‐ TMS‐related site discomfort/pain | 2 | 42 | Risk Ratio (M‐H, Fixed, 95% CI) | 8.33 [1.68, 41.27] |
| 14 Adverse effects: 2. Specific: b. Average score (CSSES, high = poor) Show forest plot | 1 | Mean Difference (IV, Fixed, 95% CI) | Subtotals only | |
| 14.1 cognitive complaints | 1 | 17 | Mean Difference (IV, Fixed, 95% CI) | ‐0.6 [‐2.69, 1.49] |
| 14.2 subjective side effects | 1 | 17 | Mean Difference (IV, Fixed, 95% CI) | ‐1.90 [‐10.31, 6.51] |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 1 Global state: Clinical improvement Show forest plot | 1 | 27 | Risk Ratio (M‐H, Fixed, 95% CI) | 4.06 [0.21, 77.37] |
| 2 Mental state: 1. General: a. Average overall mental state score (PANSS total, high = poor) Show forest plot | 3 | 108 | Mean Difference (IV, Fixed, 95% CI) | ‐5.71 [‐9.32, ‐2.10] |
| 3 Mental state: 1. General: b. Average general psychopathology score (PANSS, high = poor) Show forest plot | 3 | 108 | Mean Difference (IV, Fixed, 95% CI) | ‐2.47 [‐4.21, ‐0.73] |
| 4 Mental state: 2. Specific: a. Average negative symptom score (various scales) Show forest plot | 3 | Mean Difference (IV, Fixed, 95% CI) | Subtotals only | |
| 4.1 PANSS (high = poor) | 3 | 108 | Mean Difference (IV, Fixed, 95% CI) | ‐2.67 [‐4.25, ‐1.09] |
| 4.2 SANS (high = poor) | 1 | 27 | Mean Difference (IV, Fixed, 95% CI) | ‐11.55 [‐21.90, ‐1.20] |
| 5 Mental state: 2. Specific: b. Average positive symptom score (PANSS, high = poor) Show forest plot | 3 | 108 | Mean Difference (IV, Fixed, 95% CI) | ‐0.42 [‐1.64, 0.80] |
| 6 Cognitive state: Average score (various measures) Show forest plot | 1 | Mean Difference (IV, Fixed, 95% CI) | Subtotals only | |
| 6.1 digit span test | 1 | 39 | Mean Difference (IV, Fixed, 95% CI) | 2.10 [‐0.23, 4.43] |
| 6.2 verbal fluency test | 1 | 39 | Mean Difference (IV, Fixed, 95% CI) | 2.10 [‐2.87, 7.07] |
| 7 Adverse effects: 1. Leaving the study early Show forest plot | 2 | 76 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.36 [0.07, 1.74] |
| 8 Adverse effects: 2. Specific Show forest plot | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | Subtotals only | |
| 8.1 headache | 1 | 27 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.53 [0.11, 2.70] |
| 8.2 sleep disorder | 1 | 27 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.27 [0.01, 6.11] |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 1 Global state: Clinical improvement Show forest plot | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | Subtotals only | |
| 1.1 including only people who completed the studies | 1 | 27 | Risk Ratio (M‐H, Fixed, 95% CI) | 4.06 [0.21, 77.37] |
| 1.2 Intention‐to‐treat analysis | 1 | 30 | Risk Ratio (M‐H, Fixed, 95% CI) | 4.41 [0.23, 84.79] |
