Supresión ovárica para el tratamiento adyuvante del cáncer de mama temprano positivo para los receptores hormonales
Resumen
Antecedentes
Aproximadamente el 80% de los cánceres de mama entre las mujeres premenopáusicas son positivos para los receptores hormonales. El tratamiento endocrino adyuvante es un componente integral de la atención del cáncer de mama con receptores hormonales positivos y en las mujeres premenopáusicas incluye el bloqueo de los receptores de estrógeno con tamoxifeno, la supresión temporal de la síntesis de estrógeno ovárico mediante agonistas de la hormona liberadora de hormona luteinizante (LHRH) y la interrupción permanente de la síntesis de estrógeno ovárico mediante la ooforectomía o la radioterapia. En declaraciones de consenso internacionales recientes se recomienda el tamoxifeno o los inhibidores de la aromatasa con supresión de la función ovárica como el tratamiento endocrino adyuvante estándar actual para las mujeres premenopáusicas (a menudo precedidos de quimioterapia). Esta revisión examinó la función de agregar la SFO a otro tratamiento (es decir, quimioterapia, tratamiento endocrino o ambos), o de comparar la SFO con ningún otro tratamiento adyuvante.
Objetivos
Evaluar los efectos de la SFO para el tratamiento de mujeres premenopáusicas con cáncer de mama temprano positivo para los receptores hormonales.
Métodos de búsqueda
Para esta actualización de la revisión, se realizaron búsquedas en el Registro Especializado del Grupo Cochrane de Cáncer de Mama (Cochrane Breast Cancer Group), MEDLINE, Embase, el Registro Cochrane Central de Ensayos Controlados (Cochrane Central Register of Controlled Trials) (CENTRAL; 2019, número 8), la World Health Organization (WHO) International Clinical Trials Registry Platform (ICTRP), y ClinicalTrials.gov el 26 de septiembre 2019. Se examinaron las listas de referencias de los artículos relacionados, se estableció contacto con los autores de los ensayos y no se aplicaron restricciones de idioma.
Criterios de selección
Se incluyeron todos los ensayos aleatorizados que evaluaron cualquier método de SFO, es decir, ooforectomía, ablación ovárica inducida por radiación o agonistas de la LHRH, como tratamiento adyuvante para las mujeres premenopáusicas con cáncer de mama en estadio temprano. Se incluyeron estudios que compararon (1) SFO versus observación, (2) SFO + quimioterapia versus quimioterapia, (3) SFO + tamoxifeno versus tamoxifeno, y (4) SFO + quimioterapia + tamoxifeno versus quimioterapia + tamoxifeno.
Obtención y análisis de los datos
Dos autores de la revisión de forma independiente extrajeron los datos y evaluaron el riesgo de sesgo y la certeza de la evidencia mediante los criterios GRADE. Se obtuvieron los cociente de riesgos instantáneos (CRI) para los resultados del tiempo transcurrido hasta el evento y se realizó un metanálisis mediante un modelo de efectos fijos. Las medidas de resultado primario de interés fueron la supervivencia general (SG) y la supervivencia sin enfermedad (SSE). La toxicidad, el cáncer de mama contralateral y la segunda neoplasia maligna se representaron como riesgos relativos (RR), y se extrajeron los datos sobre la calidad de vida cuando se proporcionaron.
Resultados principales
Esta actualización de la revisión incluyó 15 estudios con 11 538 mujeres premenopáusicas con cáncer de mama temprano con receptores hormonales positivos; estos estudios se realizaron entre 1978 y 2014. Algunos de estos tratamientos no son la norma de atención actual, y los primeros estudios no evaluaron el estado de los receptores HER2. Los estudios probaron la SFO versus la observación (un estudio), la SFO más quimioterapia versus la quimioterapia (seis estudios), la SFO más tamoxifeno versus el tamoxifeno (seis estudios) y la SFO más quimioterapia y tamoxifeno versus la quimioterapia y el tamoxifeno (dos estudios). De los estudios que informaron sobre el régimen de quimioterapia, se calcula que el 72% de las mujeres recibieron una antraciclina. Los resultados que se describen a continuación se refieren a la comparación general de la SFO con ninguna SFO.
Evidencia de certeza alta muestra que el agregado de la SFO al tratamiento dio lugar a una reducción de la mortalidad (cociente de riesgos instantáneos [CRI] 0,86, intervalo de confianza [IC] del 95%: 0,78 a 0,94; 11 estudios; 10 374 mujeres; 1933 eventos informados). Este efecto del tratamiento se observó cuando se agregó la SFO a la observación, al tamoxifeno o a la quimioterapia y el tamoxifeno. El efecto sobre la mortalidad no se observó cuando la SFO se agregó a la quimioterapia sin tratamiento con tamoxifeno (CRI 0,95, IC del 95%: 0,82 a 1,09; cinco estudios; 3087 mujeres; mediana de seguimiento: rango 7,7 a 12,1 años). El agregado de la SFO dio lugar a una mejora de la SSE (CRI 0,83, IC del 95%: 0,77 a 0,90; diez estudios; 8899 mujeres; 2757 eventos informados; evidencia de certeza alta). El efecto del tratamiento sobre la SSE persistió cuando la SFO se agregó a la observación, al tamoxifeno y a la quimioterapia y el tamoxifeno. El efecto sobre la SSE se redujo cuando la SFO se agregó a la quimioterapia sin tratamiento con tamoxifeno (CRI 0,90; IC del 95%: 0,79 a 1,01; cinco estudios; 2450 mujeres). La heterogeneidad fue baja a moderada en todos los estudios para la SSE y la SG (respectivamente).
La evidencia indica que el agregado de la SFO aumenta ligeramente la incidencia de sofocos (grado 3/4 o cualquier grado; riesgo relativo [RR]1,60, IC del 95%: 1,41 a 1,82; seis estudios; 5581 mujeres; evidencia de certeza baja, ya que este resultado puede haberse informado de manera no óptima en estos estudios). Otros dos estudios que no fue posible incluir en el metanálisis informaron de un mayor número de sofocos en el grupo de SFO que en el grupo ninguna SFO. Siete estudios con 5354 mujeres recopilaron información relacionada con el estado de ánimo; sin embargo, esta información se proporcionó como depresión grado 3 o 4, ansiedad o síntomas neuropsiquiátricos, o los síntomas se informaron sin el grado. Dos estudios informaron de un aumento de la depresión, la ansiedad y los síntomas neuropsiquiátricos en el grupo de SFO en comparación con el grupo de ninguna SFO, y cinco estudios indicaron un aumento de la ansiedad en ambos grupos de tratamiento (pero ninguna diferencia entre los grupos), o ninguna diferencia general de los síntomas con el tiempo o entre los grupos de tratamiento. Un único estudio informó sobre la salud ósea como osteoporosis (definida como puntuación T < ‐2,5); esta evidencia limitada indica que la SFO aumenta el riesgo de osteoporosis en comparación con ninguna SFO en una mediana de seguimiento de 5,6 años (RR 1,16; IC del 95%: 1,10 a 28,82; 2011 mujeres; evidencia de certeza baja).
El agregado de la SFO al tratamiento probablemente reduce el riesgo de cáncer de mama contralateral (CRI 0,75; IC del 95%: 0,57 a 0,97; nueve estudios; 9138 mujeres; evidencia de certeza moderada).
La calidad de vida se evaluó en cinco estudios; en cuatro de ellos se utilizaron instrumentos validados, y en el quinto no se proporcionó información sobre la forma la en que se recopilaron los datos. Dos estudios informaron peores indicadores de calidad de vida (es decir, sequedad vaginal, sudores diurnos y nocturnos) para las mujeres que recibieron SFO, en comparación con las del grupo ninguna SFO. Los otros dos estudios indicaron un empeoramiento de los síntomas (p.ej. vasomotores, ginecológicos, sequedad vaginal, disminución del interés sexual, dolor óseo y articular, aumento de peso); sin embargo, estos efectos secundarios se informaron por igual en el grupo SFO y en el grupo ninguna SFO. El estudio que no utilizó un instrumento validado de calidad de vida no describió diferencias considerables entre los grupos.
Conclusiones de los autores
Esta revisión encontró evidencia que apoya el agregado de la SFO para las mujeres premenopáusicas con cánceres de mama tempranos y positivos para los receptores hormonales. El efecto beneficioso de la SFO persistió cuando se comparó con la observación y cuando se agregó al tratamiento endocrino (tamoxifeno) o a la quimioterapia y al tratamiento endocrino (tamoxifeno). La decisión de administrar la SFO puede depender de la evaluación general del riesgo basada en las características del tumor y de la paciente, y puede ser consecuencia de la consideración de todos los efectos secundarios que se producen con el agregado de la SFO.
PICO
Resumen en términos sencillos
Supresión de la función ovárica para el tratamiento de mujeres premenopáusicas con cáncer de mama temprano positivo para los receptores hormonales
¿Cuál era el objetivo de esta revisión?
El objetivo de esta revisión Cochrane fue determinar si el agregado de la supresión de la función ovárica al tratamiento del cáncer de mama temprano mejora la supervivencia, reduce el riesgo de reaparición del cáncer y es seguro para las mujeres premenopáusicas con cáncer de mama temprano con receptores hormonales positivos. Los autores de la revisión Cochrane recopilaron y analizaron todos los estudios relevantes para responder a estas preguntas y encontraron 15 estudios.
¿Mensajes clave?
El agregado de la supresión de la función ovárica al tratamiento mejoró la supervivencia (las mujeres vivieron más tiempo) y redujo la posibilidad de que el cáncer volviera a aparecer en mujeres con cáncer de mama operable en sus primeras etapas, pero la administración de la supresión de la función ovárica parece aumentar el riesgo de sofocos y puede afectar la salud ósea. La decisión de utilizar la SFO se debe personalizar después de considerar el perfil de riesgos y efectos beneficiosos.
¿Qué se estudió en la revisión?
Alrededor de ocho de cada diez mujeres premenopáusicas que desarrollan cáncer de mama tienen un tipo de cáncer sensible a las hormonas, llamado enfermedad con "receptor hormonal positivo". Para retrasar el crecimiento de cualquier célula cancerosa que quede después de la cirugía se puede utilizar el tratamiento hormonal para reducir la disponibilidad de los estrógenos naturales para las células cancerosas. Lo anterior se puede hacer mediante el bloqueo de los receptores de estrógeno en las células con fármacos como el tamoxifeno, la supresión de la producción de estrógeno con fármacos llamados agonistas de la hormona liberadora de hormona luteinizante (LHRH), o la extirpación de los ovarios con cirugía o al disminuir su capacidad de producir hormonas con la radioterapia.
Esta revisión examinó la función de la supresión de la función ovárica (es decir, los agonistas de la LHRH, la extracción de los ovarios o la supresión ovárica inducida por la radiación) para las mujeres premenopáusicas con cáncer de mama en estadio temprano con receptores hormonales positivos. La práctica de suprimir la función ovárica además de proporcionar otros tratamientos ha sido de interés en los últimos cinco años, a medida que se dispone de nuevos datos de los ensayos clínicos. Se necesita una revisión de estos datos para determinar los efectos beneficiosos de agregar la supresión de la función ovárica al tratamiento, identificar los efectos secundarios de la supresión de la función ovárica y descubrir cómo el tratamiento afecta el bienestar general de la mujer (calidad de vida).
La fuente de financiación para la realización de estos estudios fue el gobierno (cuatro estudios), el gobierno y las empresas farmacéuticas combinados (tres estudios), el gobierno y organizaciones sin fines de lucro combinados (dos estudios), organizaciones sin fines de lucro y empresas farmacéuticas (dos estudios), y una empresa farmacéutica (un estudio); tres estudios no informaron sobre la fuente de financiación.
¿Cuáles son los principales resultados de la revisión?
Los autores de la revisión encontraron 15 estudios relevantes que incluyeron a 11 538 mujeres. Para lograr la supresión de la función ovárica, nueve estudios administraron agonistas de la LHRH (la goserelina fue la que se administró con mayor frecuencia), dos estudios indujeron la supresión de la función ovárica a través de la cirugía y cuatro estudios permitieron cualquier método (agonistas de la LHRH, cirugía o radioterapia). Los agonistas de la LHRH se les administraron a las mujeres durante un mínimo de un año.
La salud de la mujer se monitorizó al menos durante dos años desde el comienzo del estudio. Algunos estudios monitorizaron a las mujeres durante más de 12 años.
Los autores de la revisión encontraron que agregar la supresión de la función ovárica al tratamiento:
‐ mejora la supervivencia y reduce el riesgo de que el cáncer regrese, en comparación con el tratamiento sin supresión de la función ovárica;
‐ parece aumentar las posibilidades de que se produzcan sofocos graves, en comparación con el tratamiento sin supresión de la función ovárica;
‐ probablemente reduce el riesgo de un segundo cáncer de mama en el otro seno, en comparación con el tratamiento sin supresión de la función ovárica;
‐ puede o no tener efecto sobre el estado de ánimo (p.ej. ansiedad, depresión), en comparación con el tratamiento sin supresión de la función ovárica;
‐ puede aumentar el riesgo de osteoporosis en comparación con el tratamiento sin supresión de la función ovárica (sin embargo, este hallazgo se basó en un estudio); y
‐ puede hacer poca o ninguna diferencia en cuanto a la calidad de vida de las mujeres, en comparación con el tratamiento sin supresión de la función ovárica. Cinco de 15 estudios proporcionaron información sobre la calidad de vida de las mujeres.
¿Cuán actualizada está esta revisión?
Los autores de la revisión buscaron los estudios que se habían publicado hasta septiembre 2019.
Conclusiones de los autores
Summary of findings
| OFS compared to no OFS for adjuvant treatment of early breast cancer | ||||||
| Patient or population: women with early breast cancer | ||||||
| Outcomes | Anticipated absolute effects* (95% CI) | Relative effect | № of participants | Certainty of the evidence | Comments | |
| Risk with no OFS | Risk with OFS | |||||
| Overall survival (OS) Median follow‐up: range 5.3 to 12.1 years (*baseline risk at 5 and 10 years estimated from the control arm in 8 and 7 studies, respectively) | 5‐year risk of death* | HR 0.86 | 10,374 | ⊕⊕⊕⊕ | ||
| 110 per 1000 | 95 per 1000 | |||||
| 10‐year risk of death* | ||||||
| 310 per 1000 | 273 per 1000 | |||||
| Disease‐free survival (DFS) (*baseline risk at 5 and 10 years estimated from the control arm in 9 and 7 studies, respectively) | 5‐year risk of recurrence* | HR 0.83 | 8899 | ⊕⊕⊕⊕ | ||
| 250 per 1000 | 212 per 1000 | |||||
| 10‐year risk of recurrence* | ||||||
| 440 per 1000 | 382 per 1000 | |||||
| Toxicity ‐ hot flushes (a combination of "grade 3/4" and "any grade" toxicity) Follow‐up: range 2 to 5 years | 97 per 1000 | 154 per 1000 | RR 1.60 | 5581 | ⊕⊕⊕⊝ | An additional 2 studies (studied population: 246 in ABCTCG, not detailed in IBCSG VIII) reported a higher number of hot flushes in the OFS group than in the no OFS group. These results are consistent with the overall effect estimate |
| Toxicity ‐ mood Follow‐up: range 1 year to 9.6 years | 7 out of 15 studies reported Grade 3 or 4 depression, anxiety, or neuropsychiatric symptoms or did not report the grade of these side effects. Two studies ‐ ABCTCG; E‐3193, INT‐0142 ‐ reported an increase in depression, anxiety, and neuropsychiatric symptoms in the OFS group compared to the non‐OFS group. Five studies ‐ ECOG 5188, INT‐0101; GABG IV‐B‐93; SOFT; Yi 2016; ZIPP ‐ indicated an increase in anxiety in both treatment groups (but no differences between groups) or no difference overall in symptoms over time or between treatment groups | ‐ | 5354 | ⊕⊕⊕⊝ | ||
| Toxicity ‐ bone health (osteoporosis defined by a T score < ‐2.5) Follow‐up: median 5.6 years | 35 per 1000 | 58 per 1000 | RR 1.66 (1.10 to 2.50) | 2011 | ⊕⊕⊝⊝c | |
| Contralateral breast cancer Follow‐up: 4.75 to median 12.1 years | 31 per 1000 | 23 per 1000 | RR 0.75 | 7856 | ⊕⊕⊕⊝ | |
| Quality of life | Four out of 15 studies collected data on quality of life using validated tools (ABCTCG; E‐3193, INT‐0142; IBCSG VIII; SOFT), and 1 study collected quality of life‐type information without describing a validated tool (ZBCSG Trial B). Two studies ‐ ABCTCG; E‐3193, INT‐0142 ‐ reported worse quality of life indicators (i.e. vaginal dryness, day and night sweats) in the OFS group than in the no OFS group. The other 2 studies ‐ IBCSG VIII; SOFT ‐ indicated worsening of symptoms (e.g. vasomotor, gynaecological, vaginal dryness, decline in sexual interest, bone and joint pain, weight gain); however these side effects were reported in both OFS and no OFS groups. The study that did not use a validated quality of life tool described no considerable differences between groups | ‐ | Estimated to be 2996 (5 RCTs) | ⊕⊕⊝⊝ | ||
| *The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). | ||||||
| GRADE Working Group grades of evidence. | ||||||
| aThree studies reported Grade 3 or 4 hot flushes using a standardised toxicity symptom scale (E‐3193, INT‐0142; ECOG 5188, INT‐0101; SOFT), and the other three studies that contributed to the meta‐analysis reported any grade of hot flushes/sweats without reporting the scale used (Arriagada 2005; ZBCSG Trial B; ZIPP). This outcome was downgraded due to variability in reporting of hot flushes across studies and uncertainty as to whether unblinding of treatment allocation may have affected patient reporting and concerns about selective outcome reporting. Therefore downgraded by two points overall. | ||||||
Antecedentes
Descripción de la afección
El cáncer de mama es una causa principal de morbilidad y mortalidad. Aproximadamente el 20% de las mujeres diagnosticadas y tratadas por cáncer de mama temprano finalmente morirán de la enfermedad (Jemal 2008). Según los datos de los registros de Surveillance, Epidemiology, and End Results (SEER), en los Estados Unidos aproximadamente el 80% de los tumores en mujeres premenopáusicas son sensibles a las hormonas y pueden ser adecuados para el tratamiento hormonal (Howlader 2014).
Descripción de la intervención
El objetivo del tratamiento adyuvante es prevenir la recidiva a distancia y mejorar la supervivencia general. En las mujeres premenopáusicas con cáncer de mama temprano positivo para los receptores hormonales, las opciones de tratamiento adyuvante incluyen la quimioterapia citotóxica y la terapia hormonal (Gelber 1996). El objetivo de la terapia hormonal es reducir la disponibilidad de estrógeno para las células cancerosas. Lo anterior se puede lograr al bloquear los receptores de estrógeno (p.ej. mediante el tamoxifeno) o a través de la supresión de la función ovárica (SFO) para inhibir la síntesis de estrógeno.
La SFO puede ocurrir de forma irreversible con la ooforectomía quirúrgica o con la ablación ovárica inducida por radiación. Puede ocurrir de una manera reversible con la administración de agonistas de la hormona liberadora de la hormona luteinizante (LHRH). Los agonistas LHRH actúan mediante la unión a los receptores LHRH hipofisiarios, lo que provoca una disminución de los receptores y una supresión posterior de la hormona luteinizante y el estradiol (Furr 1989). El agonista LHRH prescrito con más frecuencia es la goserelina. Los principales efectos secundarios de la SFO son la infertilidad, la disminución de la libido, los sofocos, el sudor, el dolor de cabeza, los cambios en la presión arterial, la pérdida de densidad ósea, la hipercalcemia y otras complicaciones poco frecuentes.
De qué manera podría funcionar la intervención
Se ha demostrado que la quimioterapia induce amenorrea en más del 60% al 80% de las mujeres premenopáusicas que reciben tratamiento adyuvante (Bines 1996; Walshe 2006). Las mujeres que se vuelven amenorreicas después de la quimioterapia tienen una mejor supervivencia libre de enfermedad que las que no lo hacen, en particular en el caso de las enfermedades sensibles a las hormonas (Pagani 1998; Davidson 2003; Walshe 2006). Lo anterior indica que al menos algún efecto beneficioso de la quimioterapia en las mujeres premenopáusicas está mediado por su efecto tóxico sobre los ovarios, al provocar ablación o supresión ovárica. El valor de la ablación ovárica mediante ooforectomía o radioterapia queda claramente demostrado por un metanálisis del Early Breast Cancer Trialists' Collaborative Group (EBCTCG 1996; EBCTCG 2003; EBCTCG 2005), que se centró en las mujeres menores de 50 años, la mayoría de las cuales tenían probabilidades de ser premenopáusicas. Para las mujeres que recibieron ablación ovárica a falta de quimioterapia hubo una reducción del 25% de las probabilidades anuales de recidiva, así como una reducción del 24% de las probabilidades anuales de muerte. El efecto beneficioso se observó en las mujeres con ganglios positivos y ganglios negativos. Entre las mujeres asignadas al azar a ablación ovárica después de la quimioterapia, el efecto beneficioso de la ablación pareció ser menor y no fue estadísticamente significativo (la reducción de las probabilidades anuales de recidiva fue del 10% con un error estándar [EE] del 9%; la reducción de las probabilidades anuales de muerte fue del 8%, con un EE del 10%).
Los primeros ensayos clínicos compararon los resultados de la SFO con los de la quimioterapia (SCTBG e ICRF 1993; Ejlertsen 1999); con frecuencia los investigadores no seleccionaron a las participantes en función de su estado hormonal. Las guías actuales recomiendan el tratamiento hormonal en todas las mujeres con cáncer de mama con receptores hormonales positivos (Cardoso 2019; NCCN 2019). Una de las cuestiones cada vez más importantes sobre el tratamiento de las mujeres premenopáusicas con cáncer de mama temprano ha sido el valor de agregar la SFO al tratamiento.
Por qué es importante realizar esta revisión
La función de la SFO ha sido objeto de amplias investigaciones desde 1896, incluido una revisión exhaustiva en la que se utilizaron datos de participantes individuales, realizado por el Early Breast Cancer Trialists' Collaborative Group (Clarke 1998; EBCTCG, Clarke 1998; EBCTCG 2005). Estas primeras investigaciones se realizaron antes de la era del tratamiento dirigido según el estado de los receptores hormonales, o la evaluación del estado de los receptores HER2. En recientes declaraciones de consenso internacionales se recomienda el tamoxifeno o los inhibidores de la aromatasa, con supresión de la función ovárica, como el tratamiento endocrino adyuvante estándar actual para las mujeres premenopáusicas muy jóvenes o para las mujeres premenopáusicas de alto riesgo que reciben quimioterapia. La presente revisión Cochrane examinó la función de agregar la SFO a muchos tratamientos diferentes (incluidos varios regímenes de quimioterapia, tratamiento endocrino u observación) para mujeres con cáncer de mama temprano con receptores hormonales positivos. Los hallazgos de esta revisión ayudarán a los consumidores y a los médicos, a los desarrolladores de guías y a los organismos de financiación (p.ej. NIHR UK).
El objetivo de esta revisión fue aclarar los efectos de la SFO para el tratamiento adyuvante del cáncer de mama temprano con receptores hormonales positivos en mujeres premenopáusicas, mediante una revisión sistemática de los ensayos aleatorizados disponibles. Una versión anterior de esta revisión se centra en los efectos de los agonistas de la LHRH sobre el tratamiento adyuvante del cáncer de mama con comparaciones afectadas por los factores de confusión (Goel 2009). Esta revisión sistemática actual aborda la cuestión moderna de la SFO en comparación con ninguna SFO en mujeres premenopáusicas con receptores hormonales positivos, debido a que se han informado varios ensayos completos o se han informado nuevos ensayos desde el momento en que se publicó Goel 2009, y a que los ajustes en los criterios de elegibilidad han dado lugar a la inclusión de ensayos adicionales.
Objetivos
Evaluar los efectos de la SFO para el tratamiento de mujeres premenopáusicas con cáncer de mama temprano positivo para los receptores hormonales.
Métodos
Criterios de inclusión de estudios para esta revisión
Tipos de estudios
Se incluyeron ensayos controlados aleatorizados. Los estudios pseudaleatorizados no fueron elegibles.
Se incluyeron estudios publicados como artículos de texto completo o como resúmenes de conferencias.
Tipos de participantes
Se incluyeron mujeres premenopáusicas con un diagnóstico histológico de cáncer de mama temprano, positivo para receptores hormonales. El cáncer de mama temprano se define como estadio I, II y III de la clasificación Tumor‐Ganglios‐Metástasis (TNM, por sus siglas en inglés). Los estudios definen la premenopausia como menstruación en los últimos tres a 12 meses, o niveles de estradiol en los rangos premenopáusicos.
Se excluyeron los estudios de mujeres con enfermedad metastásica.
Tipos de intervenciones
Se definió una intervención como cualquier forma de SFO (es decir, ooforectomía, ablación ovárica inducida por radiación o agonistas de la LHRH). Los agonistas de la LHRH podían incluir buserelina, goserelina, leuprorelina, nafarelina y triptorelina, y se debían administrar durante al menos 12 meses.
Se definió un comparador como cualquier régimen sin SFO. El tratamiento endocrino y la quimioterapia se permitieron si se administró el mismo tratamiento a ambos grupos.
Las comparaciones podían incluir las siguientes:
-
SFO versus observación.
-
SFO + quimioterapia versus quimioterapia.
-
SFO + tamoxifeno versus tamoxifeno.
-
SFO + quimioterapia + tamoxifeno versus quimioterapia + tamoxifeno.
Tipos de medida de resultado
Resultados primarios
-
Supervivencia general (SG), definida como el tiempo desde la fecha de asignación al azar hasta la fecha de la muerte por cualquier causa
-
Supervivencia libre de enfermedad (SLE), definida como el tiempo desde la fecha de asignación al azar hasta la primera recidiva, el cáncer de mama contralateral, el segundo cáncer de mama o la muerte, o según lo definido por el estudio
Resultados secundarios
-
Cáncer de mama contralateral
-
Segunda neoplasia maligna
-
Eventos adversos como sofocos, trastornos del estado de ánimo, reducción de la densidad ósea, artralgias, alteración de la función sexual, aumento del riesgo cardiovascular, trombosis venosa profunda, embolia pulmonar, deterioro de la función cognitiva, muerte relacionada con el tratamiento y cualquier otra toxicidad importante, informados en los estudios. Las toxicidades se podían definir según los criterios de toxicidad World Health Organization (WHO)/National Cancer Institute of Canada (NCIC), o según el estudio
-
Cumplimiento con el tratamiento
-
Calidad de vida (CdV), evaluada mediante instrumentos validados o específicos para los ensayos, como el European Organisation for Research and Treatment of Cancer (EORTC) Quality of Life Questionnaire
Métodos de búsqueda para la identificación de los estudios
Búsquedas electrónicas
We searched the following databases on 17 September 2018 and performed a top‐up search on 26 September 2019.
-
Specialised Register of the Cochrane Breast Cancer Group. Details of the search strategy used by the Group for identification of studies and the procedure used to code references are outlined in the Group's module (www.mrw.interscience.wiley.com/cochrane/clabout/articles/BREASTCA/frame.html). We carried out a search for the following text words: 'buserelin', 'goserelin', 'gonadotropin‐releasing hormone', 'leuprorelin', 'triptorelin', 'nafarelin', 'LHRH', 'oophorectomy', 'ovariectomy', 'ablation', and 'ovarian function depression'.
-
Cochrane Central Register of Controlled Trials (CENTRAL; 2019, Issue 8) (see Appendix 1).
-
MEDLINE (via OvidSP) (see Appendix 2).
-
Embase (via OvidSP) (see Appendix 3).
-
World Health Organization (WHO) International Clinical Trials Registry Platform (ICTRP) search portal for all prospectively registered and ongoing trials (see Appendix 4).
-
ClinicalTrials.gov register (clinicaltrials.gov) for additional unpublished and ongoing studies (see Appendix 5).
Búsqueda de otros recursos
We contacted some of the investigators of potentially eligible studies for unpublished data or clarification of data analysis (i.e. whether or not analyses were adjusted) and checked PubMed to learn whether eligible conference abstracts had been published as full‐text articles. These approaches are recognised as appropriate methods (Young 2011; Scherer 2018).
For the previous versions of this review (Goel 2009), we handsearched the proceedings of annual meetings of the American Society of Clinical Oncology and the San Antonio Breast Cancer Symposium (2005 to 2008). For the 2019 review update, handsearching of these conference proceedings was not required because these are now imported and searched via the aforementioned Embase search (as outlined in Appendix 3).
Obtención y análisis de los datos
Selección de los estudios
We applied the eligibility criteria to each of the retrieved references. In the first instance, we used study publications to assess each study's eligibility. If a study had not been published, we attempted to find the necessary information from a study protocol or a clinical trial registry record.
For the original review and the review update, two review authors (review update: TB, AG) independently assessed each potentially eligible study. A third review author was not required as there were no disagreements regarding eligibility.
We recorded excluded studies in the Characteristics of excluded studies table.
We applied no language restrictions.
Extracción y manejo de los datos
For the original review and the review update, two review authors (review update: TB, MW)) independently extracted data from the included studies. If required, a third review author (AG) was available to resolve any discrepancies regarding extraction of quantitative data. We collected information on study design, participants (including hormone receptor status and nodal involvement), settings, interventions, primary and secondary outcomes, follow‐up, and sources of funding. For studies with more than one publication, we extracted data from these publications, and we considered the final or updated version of each study as the primary reference. For one included study, four colleagues at the Japanese Cochrane Centre conducted data extraction and risk of bias assessments and double‐checked the translated material.
Evaluación del riesgo de sesgo de los estudios incluidos
For the review update, we used Cochrane's 'Risk of bias' assessment tool to assess potential sources of bias in the included studies (Higgins 2011). Two review authors (TB, MW) independently assessed the potential risk of bias for each study and resolved any differences in judgement through discussion. The domains assessed were random sequence generation, allocation concealment, blinding of participants and personnel, blinding of outcome assessment, incomplete outcome data, selective reporting, and other bias. We assigned a rating of 'high', 'low', or 'unclear' risk of bias to each domain for each included study in keeping with the criteria outlined in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011). Among phase 3 oncology studies, open‐label studies are common due to the difficulty involved in concealing different chemotherapy schedules and toxicities. We therefore grouped the blinding of outcome assessment domain with outcome measures from most unlikely to most likely to be influenced by lack of blinding. We segregated outcomes into (1) OS, (2) DFS, (3) toxicity, and (4) quality of life.
Medidas del efecto del tratamiento
Two review authors extracted data from each trial.
The primary outcomes for this review were OS and DFS, with both considered as time‐to‐event outcomes. Hazard ratios (HRs) and variances were extracted from trial publications, when available. If not reported, statistics were extracted from publications via the methods described by Tierney et al using other summary statistics (Tierney 2007). These indirect methods were recorded in the Notes section in the Characteristics of included studies tables. All efficacy analyses used an intention‐to‐treat population when this was reported. Furthermore, data for OS and DFS were extracted for the hormone receptor‐positive population or if more than 50% of the study population had hormone receptor‐positive cancers.
Contralateral breast cancer and second malignancy were collected and reported as risk ratios (RRs) with 95% confidence intervals (CIs).
Toxicity data were extracted from each study; when possible, this was done for the treated population rather than the intention‐to‐treat population. As definitions of toxic events varied between trials, events were extracted and summarised to best reflect clinically important outcomes. Pooled RRs and 95% CIs were calculated for hot flushes; all other toxicities were presented as frequencies and proportions and were not included as part of a meta‐analysis due to the paucity of data. For this review update, if efficacy data were reported separately for hormone receptor‐positive cancers but not for toxicity outcomes, we extracted the toxicity data regardless of the proportion of the studied population with hormone receptor‐positive tumours because we expected toxicity to be the same regardless of hormone receptor status.
We collected quality of life data irrespective of the questionnaire used. We made no attempt to statistically synthesise quality of life data, which we summarised and reported qualitatively.
Cuestiones relativas a la unidad de análisis
Three studies were three‐arm studies (ECOG 5188, INT‐0101; IBCSG VIII; ZBCSG Trial B). For all three studies, data from two of the three arms were used and were relevant for this review topic. The third arm contained a confounded comparator or intervention group.
One study was a 2 × 2 factorial study (ZIPP). For the analysis, there were two relevant intervention arms (goserelin ± elective tamoxifen, and goserelin + randomised tamoxifen) and two relevant comparator arms (observation ± elective tamoxifen, and tamoxifen). Data from the two intervention arms were combined and were compared to data from the two comparator arms.
Manejo de los datos faltantes
When data were missing, we contacted the original investigators (by written correspondence) to request missing data. For the review update, we contacted the following trialists for summary statistics, numbers of events for each treatment arm (for overall survival or disease‐free survival), and clarification on whether HRs were adjusted or unadjusted: ABCTCG; Arriagada 2005; ASTRRA; GABG IV‐B‐93; Uslu 2014; Yang 2013. We received additional data from the trialists for two studies: ASTRRA; GABG IV‐B‐93.
Evaluación de la heterogeneidad
We assessed heterogeneity by using the Chi² test and the I² statistic and by visually inspecting forest plots. We inspected the graphical representation of data; if confidence intervals for the results of individual studies had poor overlap, this generally indicated the presence of statistical heterogeneity.
We interpreted the I² statistic as per guidance provided in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011): 0% to 40% might not be important; 30% to 60% represented moderate heterogeneity; 50% to 90% represented substantial heterogeneity; and 75% to 100% represented considerable heterogeneity.
As there was minimal heterogeneity in the majority of the studies analysed in this review, we used the fixed‐effect model. When there was considerable heterogeneity (for hot flushes), we used the random‐effects model and explored sources of heterogeneity; however we ultimately used a fixed‐effect model for these given that the conclusions were the same based on fixed‐effect and random‐effects analyses.
Evaluación de los sesgos de notificación
We followed recommendations for testing for funnel plot asymmetry as described in Section 10.4.3.1 of the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011). Funnel plot asymmetry may be due to reporting bias; we addressed this possibility in the Results and Discussion sections of the review for the two primary outcomes. Supplementary to visual inspection of the funnel plot, we conducted Egger's test for the primary outcomes using R (metafor package; R).
Síntesis de los datos
For time‐to‐event outcome data (i.e. OS and DFS), we used a fixed‐effect (inverse‐variance method) model.
For dichotomous outcome data (i.e. contralateral breast cancer and second malignancy), we used the fixed‐effect (inverse‐variance method) model. For the toxicity outcome ‐ hot flushes ‐ we used the fixed‐effect (Mantel‐Haenszel method) model. For all other outcomes (including most toxicity outcomes and compliance with treatment), we reported data when available and summarised the information narratively.
We performed all analyses using Review Manager software (RevMan).
Summary of findings
We used the GRADE approach to assess the certainty of evidence for the following seven main outcomes: overall survival, disease‐free survival, hot flushes (grade 3/4), mood, bone health, contralateral breast cancer, and quality of life. We used GRADEproGDT software to develop the 'Summary of findings' table and followed GRADE guidance (GRADEproGDT; Schünemann 2019). Two review authors (TB, MW) graded the certainty of evidence for this review update.
To calculate absolute risk of the comparator group for time‐to‐event outcomes, we estimated the event rate at two specific time points (i.e. five and ten years for overall survival and disease‐free survival) from the Kaplan‐Meier curves or reported event rates. We entered these estimated values into GRADEproGDT, and the corresponding absolute risks for the intervention group at five and ten years were automatically populated by GRADEproGDT.
Análisis de subgrupos e investigación de la heterogeneidad
We planned to perform the following post‐hoc subgroup analysis for OS and DFS: HER2 ISH status: positive or negative; age group: younger than 35 years of age versus 35 to 40 years of age versus over 40 years of age; molecular subtypes: luminal A versus luminal B; chemotherapy regimen: non‐anthracycline/taxane versus anthracycline/taxane versus dose‐dense anthracycline/taxane; breast cancer stage: locally advanced breast cancer that is inoperable at presentation (stage III) versus stage I/II breast cancer at presentation. However, data were not available for these analyses.
We performed the following post‐hoc subgroup analyses for OS and DFS.
-
Duration of OFS: fewer than three years versus three years or longer.
-
Age of studied population: younger than 40 years versus 40 years of age or older.
-
Chemotherapy use irrespective of treatment combinations (i.e. chemotherapy alone or with endocrine therapy): yes or no.
-
Method of OFS: surgery versus LHRH agonists versus radiation‐induced ovarian ablation.
-
Lymph node status: positive (defined as ≥ 50% of population with node‐positive disease) versus negative (defined as < 50% of the population with node‐positive disease).
Análisis de sensibilidad
We planned to perform sensitivity analysis in relation to studies that were at high risk of bias and publication status (fully published trials versus trials published in abstract form only). However, none of the included studies met these criteria; therefore sensitivity analyses were not conducted.
Results
Description of studies
Results of the search
For this review update, searching yielded 1987 records from the Specialised Register of the Cochrane Breast Cancer Group, MEDLINE, Embase, CENTRAL, WHO ICTRP, and ClinicalTrials.gov on 26 September 2019. Searching relevant review papers revealed one additional record. After removing duplicates, we screened the titles and abstracts of the 1846 remaining records and excluded 1697 of them based on information found in the abstract alone. We further assessed the full‐text articles or ongoing study records for 149 records. We excluded 86 records after full‐text review and provided reasons in the PRISMA flowchart (Figure 1). Further details for exclusion are provided for a subset of studies in the Characteristics of excluded studies table.
Study flow diagram: 2019 review update.
Of the 63 remaining records, 62 records related to 15 eligible studies (ABCTCG; Arriagada 2005; ASTRRA; E‐3193, INT‐0142; ECOG 5188, INT‐0101; GABG IV‐B‐93; IBCSG II; IBCSG VIII; SOFT; SWOG 1996; Uslu 2014; Yang 2013; Yi 2016; ZBCSG Trial B; ZIPP), and one record was classified as an 'ongoing' study (NCT02132390). An additional record relating to an already identified eligible study was noted in a supplementary PubMed search. In sum, upon applying eligibility criteria, we identified 15 eligible randomised trials (relating to 64 records) that addressed the role of OFS in the adjuvant treatment of premenopausal women with hormone receptor‐positive early breast cancer. One relevant ongoing trial was identified and described in the Characteristics of ongoing studies table.
We have outlined the search process in the PRISMA flowchart in Figure 1. Details of the PRISMA flowchart for the previous version of this review can be found in Goel 2009.
Included studies
See the Characteristics of included studies table.
The 15 included studies, involving 11,538 premenopausal women with hormone‐positive early breast cancers, contributed to the following treatment comparisons.
-
OFS versus observation: one study (ZIPP).
-
OFS + chemotherapy versus chemotherapy: six studies (Arriagada 2005; ECOG 5188, INT‐0101; GABG IV‐B‐93; IBCSG II; IBCSG VIII; SWOG 1996).
-
OFS + tamoxifen versus tamoxifen: six studies (ABCTCG; E‐3193, INT‐0142; SOFT; Yang 2013; Yi 2016; ZBCSG Trial B), although the ZIPP study would have been eligible if long‐term data had been reported separately for this comparison.
-
OFS + chemotherapy + tamoxifen versus chemotherapy + tamoxifen: two studies (ASTRRA; Uslu 2014).
Standard dosing of LHRH agonists was observed across studies (i.e. goserelin 3.6 mg subcutaneously, triptorelin 3.75 mg intramuscularly, leuprorelin 3.75 mg subcutaneously, every 4 weeks). When tamoxifen was used, the most common dose was 20 mg daily (ABCTCG; ASTRRA; E‐3193, INT‐0142; SOFT; Uslu 2014). One study allowed 20 mg or 40 mg daily (ZIPP), another study used tamoxifen 10 mg twice daily (Yang 2013), one study allowed tamoxifen 10 mg twice daily or 20 mg daily (ZBCSG Trial B), and another study did not specify the dosing regimen used (Yi 2016).
Ovarian function suppression versus observation
ZIPP was an international collaboration between four breast cancer research groups that adopted similar protocols with the intention of combining their results in a prospective meta‐analysis. The study used a 2 × 2 factorial design to randomise 2710 participants into four arms of goserelin and tamoxifen, goserelin, tamoxifen alone, and observation alone. Study medications were continued for two years. Elective tamoxifen was allowed in two of the four collaborative groups. Fifty‐one per cent of participants had cancers that were oestrogen‐positive, and 53% were node‐positive, and 43% of participants received (neo)adjuvant chemotherapy. Overall, 48% of participants electively received tamoxifen, which included 95% of participants in whom it was permissible to do so (Baum 2006). The median 12‐year follow‐up data from the ZIPP collaboration included complete efficacy outcomes only for the comparison of OFS or no OFS in the overall population, including both hormone receptor‐positive and hormone receptor‐negative cancers. Results that were stratified by hormone status were reported in subgroups by the receipt of tamoxifen (electively or by randomisation). For this reason, efficacy outcomes for the ZIPP collaboration have been reported under the OFS compared to observation analysis only; it is the only study that performed this comparison.
Ovarian function suppression + chemotherapy versus chemotherapy
These six studies randomised 4376 women to OFS and chemotherapy compared to chemotherapy alone (Arriagada 2005; ECOG 5188, INT‐0101; GABG IV‐B‐93; IBCSG II; IBCSG VIII; SWOG 1996). Four studies recruited participants with mostly hormone receptor‐positive cancers (100% in ECOG 5188, INT‐0101 and SWOG 1996; 90% in IBCSG VIII; 70% in Arriagada 2005), and only 33% and 40% of participants had hormone receptor‐positive cancers in IBCSG II and GABG IV‐B‐93, respectively. Most participants had node‐positive disease (100% in IBCSG II, ECOG 5188, INT‐0101, and SWOG 1996, and at least 90% in Arriagada 2005 and GABG IV‐B‐93) with the exception of one study, which recruited only participants with node‐negative disease (IBCSG VIII). Five studies mandated the type of chemotherapy provided, although the regimen varied between studies. Mandated chemotherapy regimens included cyclophosphamide, methotrexate, and fluorouracil (CMF); CMF and prednisone; CMF, vincristine, and prednisone; cyclophosphamide, doxorubicin, and fluorouracil; and epirubicin and cyclophosphamide followed by CMF. Delivery of OFS also varied. Three studies used goserelin: IBCSG VIII (18 months), GABG IV‐B‐93 (two years), and ECOG 5188, INT‐0101 (five years). Two studies used oophorectomy (IBCSG II; SWOG 1996), and one study allowed any method of OFS (oophorectomy, pelvic radiotherapy, or triptorelin for three years; Arriagada 2005).
Ovarian function suppression + tamoxifen versus tamoxifen
These six studies randomised 3504 women to receive ovarian function suppression and tamoxifen compared to tamoxifen alone (ABCTCG; E‐3193, INT‐0142; SOFT; Yang 2013; Yi 2016; ZBCSG Trial B). Five studies included only participants with hormone receptor‐positive cancers (E‐3193, INT‐0142; SOFT; Yang 2013; Yi 2016; ZBCSG Trial B), and only 39% of participants had hormone receptor‐positive cancers in ABCTCG. Node positivity varied from 0% to 61%. Chemotherapy was permitted in three studies (ABCTCG; SOFT; Yang 2013), with rates varying from 53% to 88%. Chemotherapy use was not allowed in one study (E‐3193, INT‐0142), and chemotherapy use was not specified in two studies (Yi 2016; ZBCSG Trial B). Delivery OFS varied. All studies allowed LHRH agonists to be used, although one study used triptorelin (SOFT), three studies used goserelin (Yang 2013; Yi 2016; ZBCSG Trial B), and two studies allowed either goserelin or leuprorelin (ABCTCG; E‐3193, INT‐0142). Two studies allowed other methods of OFS (oophorectomy, pelvic radiotherapy; ABCTCG; E‐3193, INT‐0142). The duration of LHRH agonists varied between 12 months (Yi 2016), 18 months (Yang 2013), 2 years (ABCTCG; ZBCSG Trial B), and 5 years (E‐3193, INT‐0142; SOFT). The duration of tamoxifen therapy varied between 12 months (Yi 2016), 2 years (ZBCSG Trial B), and 5 years (ABCTCG; E‐3193, INT‐0142; SOFT; Yang 2013).
Ovarian function suppression + chemotherapy + tamoxifen versus chemotherapy + tamoxifen
These two studies randomised 1390 women to receive OFS, chemotherapy, and tamoxifen compared to chemotherapy and tamoxifen (ASTRRA; Uslu 2014). All participants had hormone receptor‐positive cancers. All cancers were node positive in Uslu 2014, and 56% of cancers were node positive in ASTRRA. In both studies, an anthracycline‐based chemotherapy was predominantly used (Uslu 2014 100%; ASTRRA 94%), tamoxifen was continued for five years, and OFS was achieved with goserelin for two years. Uslu 2014 allowed a switch to an aromatase inhibitor if menopause occurred whilst on tamoxifen alone.
Excluded studies
We excluded 86 records from this review update and provided a list of notable excluded studies under Characteristics of excluded studies. The main reason for excluding these studies was use of an incorrect comparator where the comparator arm received additional or different treatment regimens from the intervention arm and therefore was confounded (ABCSG 5; ABCSG‐12; Baum 1996; FASG 06; GABG IV‐A‐93; Grocta 02; Li 2019; MAM 01 GOCSI; PERCHE; Ragaz 1997; Soreide 2002; TABLE; Yu 2019; ZEBRA). In addition, one study stratified participants by oophorectomy status and did not randomise to either OFS or no OFS (Manson 2019), one study did not report outcomes by hormone receptor status (Pretoria), one study used either LHRH or tamoxifen and did not report data separately for LHRH (HMFEC), and two studies were registered or published in the 1990s or early 2000s and no further details have been published since that time (Baum 1996; UKCCR).
Risk of bias in included studies
Refer to Figure 2 for a summary of risk of bias judgements for the included studies for each risk of bias domain.
Risk of bias summary: review authors' judgements about each risk of bias item for each included study.
Allocation
The 15 studies were described as randomised. The method of random sequence generation was described adequately (i.e. with low risk of bias) in 12 studies (ABCTCG; Arriagada 2005; ASTRRA; E‐3193, INT‐0142; ECOG 5188, INT‐0101; GABG IV‐B‐93; IBCSG II; IBCSG VIII; SOFT; SWOG 1996; Yang 2013; ZIPP). These studies reportedly stratified randomisation or permuted block design and/or had no baseline imbalances. It was not possible to accurately assess the method of random sequence generation used in three studies owing to lack of information presented in the published article or the presence of imbalanced randomised arms (Uslu 2014; Yi 2016; ZBCSG Trial B). In particular, ZBCSG Trial B recruited 20 participants in the OFS group and 94 participants in the comparator group. These studies were classified as having unclear risk of bias.
Ten of the 15 studies were at low risk of bias for allocation concealment. These studies described central randomisation systems (internet‐based or co‐ordinating centre) (ABCTCG; Arriagada 2005; ASTRRA; E‐3193, INT‐0142; GABG IV‐B‐93; IBCSG II; IBCSG VIII; SOFT; Uslu 2014; ZIPP). Five studies did not describe methods of allocation concealment used or did not provide sufficient detail in the trial publication and were judged as having unclear risk of bias (ECOG 5188, INT‐0101; SWOG 1996; Yang 2013; Yi 2016; ZBCSG Trial B).
Blinding
All fifteen studies were described as 'open‐label' or most likely were open‐label studies, but this was not specifically mentioned in the trial publication. Performance bias was not considered to be a concern given that there was considerable equipoise at the time at which these studies were conducted such that knowing the treatment allocation was unlikely to affect the behaviour of clinicians and participants. Therefore we judged the 15 studies as having low risk of bias for this domain.
Detection bias was assessed by grouping outcomes with similar risks of bias: (1) OS, (2) DFS, (3) toxicity, and (4) quality of life. For OS and DFS, lack of blinding was perceived as unlikely to have an impact on this outcome assessment. Therefore all studies that reported these outcomes were perceived to be at low risk of bias. For outcome measures that were more likely to be influenced by lack of blinding (i.e. toxicity), we assessed whether outcome assessments were made using validated and standardised grading of symptom assessment tools and included biochemical tests. Of the 12 studies that reported toxicities, five used standardised grading symptoms, and therefore knowing the treatment allocation may have had an immaterial effect on the grading of symptoms by clinicians or participants. These five studies were rated as having low risk of bias on this domain (E‐3193, INT‐0142; IBCSG II; IBCSG VIII; SOFT; SWOG 1996). For the other seven studies, toxicities were self‐reported with no standardised tools; therefore reporting of this outcome may have been affected by lack of blinding. These seven studies were rated as having unclear risk of bias (ABCTCG; Arriagada 2005; ECOG 5188, INT‐0101; GABG IV‐B‐93; Yi 2016; ZBCSG Trial B; ZIPP). Quality of life measures were likely to be affected by lack of blinding to treatment. Five studies planned to collect and report quality of life (QoL) data using validated questionnaires (ABCTCG; E‐3193, INT‐0142; IBCSG VIII; SOFT; ZBCSG Trial B). Quality of life questionnaires were completed by participants who were unblinded to the treatment allocation; therefore it is uncertain whether this introduced risk of bias. We judged these five studies as having unclear risk of bias for this domain.
Incomplete outcome data
Thirteen studies described intention‐to‐treat analysis and minimal patient loss to follow‐up that was accounted for; therefore we judged them to be at low risk of bias: ABCTCG; Arriagada 2005; ASTRRA; E‐3193, INT‐0142; GABG IV‐B‐93; IBCSG II; IBCSG VIII; SOFT; SWOG 1996; Uslu 2014; Yi 2016; ZBCSG Trial B; ZIPP. One study was judged as having unclear risk of bias due to reporting that the data were analysed as intention‐to‐treat; however 34 participants were excluded from the analysis, and the division of excluded participants across treatment allocations was not provided in the trial publication (ECOG 5188, INT‐0101). The trial publication did state that results were similar between the modified intention‐to‐treat analysis and the full intention‐to‐treat analysis but did not report details. One study was judged as having high risk of bias because study authors did not analyse data by intention‐to‐treat and stated that 19 participants dropped out but provided no reasons or mention of the split across treatment groups (Yang 2013).
Selective reporting
Seven studies reported results for outcomes listed in the methods section of the trial publication (E‐3193, INT‐0142; ECOG 5188, INT‐0101; SOFT; Uslu 2014; Yi 2016; ZIPP) or provided a trial registration record with listed outcomes found in the methods and results sections of the trial publication (SOFT). In seven studies, it was assessed that there was partial or no reporting of toxicity outcomes when it was very likely that these outcomes were collected (ABCTCG; Arriagada 2005; ASTRRA; GABG IV‐B‐93; IBCSG II; ZBCSG Trial B), or that only partial numerical data were provided (IBCSG VIII). In the case of the ASTRRA study, previous publications had reported that data relating to tolerability of the medicines would be assessed. Therefore these seven studies were ranked as having unclear risk of bias. Yang 2013 was judged as having high risk of bias for this domain because not all outcomes measured (i.e. toxicities) were reported in the trial publication, and new outcomes were re‐assigned as primary outcomes in the final trial publication but were not reported in previous trial publications nor in the clinical trial registry record.
Other potential sources of bias
No other sources of bias were evident in 15 studies (ABCTCG; Arriagada 2005; ASTRRA; E‐3193, INT‐0142; ECOG 5188, INT‐0101; GABG IV‐B‐93; IBCSG II; IBCSG VIII; SOFT; SWOG 1996; Uslu 2014; Yang 2013; Yi 2016; ZBCSG Trial B; ZIPP).
Effects of interventions
Refer to summary of findings Table for the main comparison.
Overall survival
Thirteen of 15 studies collected overall survival data (all except Yi 2016 and ZBCSG Trial B); however, two studies did not provide sufficient information for analysis (IBCSG II; Uslu 2014). With median follow‐up across studies ranging from 5.3 to 12.1 years, the remaining 11 studies showed that adding OFS to treatment resulted in a reduction in mortality (HR 0.86, 95% CI 0.78 to 0.94; high‐certainty evidence; Figure 3). A total of 10,374 women were included in the OS analysis, with an estimated 1933 deaths reported from nine of the eleven studies (neither ABCTCG nor GABG IV‐B‐93 provided the number of events in each treatment group). A funnel plot and Egger's test did not support any publication bias for the studies reviewed (Figure 4; Egger's test: P = 0.12).
Forest plot of comparison: 1 OFS versus no OFS, outcome: 1.1 Overall survival.
Funnel plot of comparison: 1 OFS versus no OFS, outcome: 1.1 Overall survival.
This treatment effect was present when OFS was added to observation (HR 0.83, 95% CI 0.71 to 0.97; 1 study; 2706 women), to tamoxifen (HR 0.74, 95% CI 0.59 to 0.93; 4 studies; 3299 women), or to chemotherapy and tamoxifen (HR 0.31, 95% CI 0.10 to 0.94; 1 study; 1282 women). The effect was reduced when ovarian function suppression was added to chemotherapy only (HR 0.95, 95% CI 0.82 to 1.09; 5 studies; 3087 women). Refer to Analysis 1.1.
Subgroup analysis
Duration of OFS
The addition of OFS did not result in reduced mortality when OFS was used for three years or longer (HR 0.93, 95% CI 0.81 to 1.07; 5 studies; 4580 participants); however the studies contributing to this result had provided follow‐up for a period ranging from 7.7 to 9.89 years. The effect of OFS during adjuvant treatment resulted in a reduction in mortality when OFS was used for less than three years but for longer than one year (HR 0.79, 85% CI 0.69 to 0.91; 5 studies; 4956 women). For these studies, women had been monitored for a longer time period, with most studies providing a median follow‐up period of 12 years.
Age of studied population
Two studies reported overall survival by age and were underpowered (women younger than 40 years of age: HR 0.73, 95% CI 0.51 to 1.04; 394 women; women who were 40 years of age or older: HR 0.89, 95% CI 0.69 to 1.14; 1175 women) (ECOG 5188, INT‐0101; IBCSG VIII). Refer to Analysis 3.1
Chemotherapy use irrespective of treatment combinations
Two studies in which chemotherapy was not mandatory reported overall survival by receipt of chemotherapy (SOFT; ZIPP). In all studies that used chemotherapy, the addition of OFS resulted in a reduction in mortality (HR 0.86, 95% CI 0.76 to 0.97; 8 studies; 5453 women), and in those studies in which women did not receive chemotherapy, the addition of OFS did not reduce mortality (HR 0.89, 95% CI 0.62 to 1.28; 3 studies; 1286 women). Refer to Analysis 4.1.
Method of OFS
Reduction in mortality with the addition of OFS was observed when LHRH agonists were used for OFS (HR 0.80, 95% CI 0.71 to 0.89; 8 studies; 8101 women). No mortality benefit was noted when OFS was provided via surgery (HR 0.86, 95% CI 0.57 to 1.28, 2 studies; 415 women) or radiotherapy (HR 1.75, 95% CI 0.50 to 6.16; 1 study; 77 women). Refer to Analysis 5.1.
Lymph node status
Reduction in mortality was observed with the addition of OFS in lymph node‐positive cancers (HR 0.89, 95% CI 0.81 to 0.99; 7 studies; 7340 women) and in lymph node‐negative cancers (HR 0.69, 95% CI 0.53 to 0.88; 3 studies; 2943 women). Refer to Analysis 6.1.
Disease‐free survival
Thirteen of the 15 studies collected DFS data (all except Yi 2016 and ZBCSG Trial B); however, three studies did not provide sufficient information for analysis (ABCTCG; IBCSG II; Uslu 2014). IBCSG II did not report outcomes by hormonal status in the follow‐up publication. Median follow‐up ranged from 4 to 12.1 years. The addition of OFS resulted in improvement in DFS (HR 0.83, 95% CI 0.77 to 0.90; 10 studies; high‐certainty evidence; Figure 5). A total of 8899 women were included in the DFS analysis, with an estimated 2757 DFS events reported from nine studies (Yang 2013 did not provide the number of events in each treatment group). A funnel plot did not support any publication bias for the studies reviewed (Figure 6; Egger's test: P = 0.6285).
Forest plot of comparison: 1 OFS versus no OFS, outcome: 1.2 Disease‐free survival.
Funnel plot of comparison: 1 OFS versus no OFS, outcome: 1.2 Disease‐free survival.
The effect persisted when OFS was added to observation (HR 0.82, 95% CI 0.73 to 0.93; 1 study; 2706 women), to tamoxifen (HR 0.76, 95% CI 0.76 to 0.92; 3 studies; 2461 women), and to chemotherapy and tamoxifen (HR 0.69, 95% CI 0.48 to 0.99; 1 study; 1282 women). The effect was reduced when OFS was added to chemotherapy only (HR 0.90, 95% CI 0.79 to 1.01; 5 studies; 2450 women). Refer to Analysis 1.2
Subgroup analysis
Duration of OFS
The addition of OFS improved DFS among participants regardless of whether OFS was continued for three years or longer (HR 0.88, 95% CI 0.78 to 0.98; 5 studies; 3943 women) or for less than three years (HR 0.80, 95% CI 0.72 to 0.88; 5 studies; 4956 women; Analysis 2.2). In the three studies that reported DFS by age, a large improvement in disease‐free survival was seen with the addition of OFS in women younger than 40 years of age (HR 0.65, 95% CI 0.50 to 0.83; 3 studies; 1764 women), and no difference was seen among women 40 years of age or older (HR 0.95, 95% CI 0.78 to 1.15; 3 studies; 1504 women). Refer to Analysis 3.2.
Chemotherapy use irrespective of treatment combinations
In studies in which chemotherapy was not mandatory, two studies reported DFS by receipt of chemotherapy (SOFTZIPP). In all studies that used chemotherapy, the addition of OFS resulted in improvement in disease‐free survival among women who received chemotherapy (HR 0.86; 95% CI 0.76 to 0.97; 8 studies; 5453 women) but not among women who did not receive chemotherapy (HR 0.87, 95% CI 0.62 to 1.28; 3 studies). Refer to Analysis 4.2.
Method of OFS
Improvement in DFS was seen with the addition of OFS when LHRH agonists were used as the OFS method (HR 0.81, 95% CI 0.75 to 0.88; 8 studies; 8101 women), but not when the OFS method was surgery (HR 0.96, 95% CI 0.70 to 1.30; 2 studies; 415 women) or radiotherapy (HR 0.94, 95% CI 0.28 to 3.13; 1 study; 77 women), although the certainty of evidence for this is considered to be low due to small sample size and wide confidence intervals. Refer to Analysis 5.2.
Lymph node status
The addition of OFS resulted in improvement in DFS in lymph node‐positive cancers (HR 0.86, 95% CI 0.79 to 0.93; 6 studies; 5865 women) and in lymph node‐negative cancers (HR 0.75, 95% CI 0.64 to 0.89; 2943 women). Refer to Analysis 6.2.
Contralateral breast cancer
Nine of 15 studies reported outcomes of contralateral breast cancer (Arriagada 2005; ASTRRA; ECOG 5188, INT‐0101; GABG IV‐B‐93; IBCSG VIII; SOFT; SWOG 1996; Uslu 2014; ZIPP). The addition of OFS likely reduces the risk of contralateral breast cancer (HR 0.75, 95% CI 0.57 to 0.97; moderate‐certainty evidence; Analysis 1.3). A total of 9138 women were included in this analysis, with an estimated 196 contralateral breast cancers reported during the follow‐up period (range 4 to 10 years). Data were insufficient to meaningfully report on effects of OFS on contralateral breast cancer outcomes by duration of OFS, age, type of OFS, receipt of chemotherapy, or lymph node status.
Second malignancy
Seven of 15 studies reported outcomes of second malignancy (Arriagada 2005; ASTRRA; ECOG 5188, INT‐0101; GABG IV‐B‐93; IBCSG VIII; SOFT; SWOG 1996). The addition of OFS likely does not reduce the risk of second malignancy (HR 0.89, 95% CI 0.64 to 1.25; moderate‐certainty evidence; 6327 women). Data were insufficient to meaningfully report on effects of OFS on contralateral breast cancer outcomes by duration of OFS, age, type of OFS, receipt of chemotherapy, or lymph node status.
Hot flushes
Eight of 15 studies reported on hot flushes/sweats, with six of these studies contributing to a pooled analysis. Evidence suggests that the addition of OFS slightly increases the incidence of hot flushes (grade 3/4 or any grade; risk ratio (RR) 1.60, 95% CI 1.41 to 1.82; low‐certainty evidence; Analysis 1.5; Figure 7). Two additional studies reported a higher number of hot flushes in the OFS group compared to the not OFS group (ABCTCG; IBCSG VIII). Refer to Table 1.
Forest plot of comparison: 1 OFS versus no OFS, outcome: 1.5 Hot flushes.
| Study | Hot flushes/menopausal symptoms# | Bone health: bone density/fractures | Arthralgias (joint pain) | Mood (anxiety, depression, other as indicated in footnotes) | ||||
| OFS (n/N) | Comparator (n/N) | OFS (n/N) | Comparator (n/N) | OFS (n/N) | Comparator (n/N) | OFS (n/N) | Comparator (n/N) | |
| OFS vs observation | ||||||||
| ZIPP* | Reported as: | Reported as: | NR | NR | 17/591 | 6/600 | 34/591 | 11/600 Reported as anxiety/depression/irritability |
| OFS + tamoxifen vs tamoxifen | ||||||||
| Participants in the OFS group (n = 118) experienced more menopausal symptoms (night sweats, P = 0.005) and day sweats (P < 0.001) than those in the no OFS group (n = 128) | NR | NR | NR | NR | Participants in the OFS group had increased depression (P = 0.05) and anxiety (P = 0.04) over 30 months compared to those in the no OFS group (total number of participants = 436, data not reported separately by treatment group) | |||
| 28/174 | 8/171 | NR | NR | NR | NR | 4b/174 | 4b/171 | |
| 133/1005 | 76/1006 | 3/1005 | 1/1006 | 55/1005 | 63/1006 | 44/1005 | 38/1006 | |
| NR | NR | NR | NR | NR | NR | NR | NR | |
| NR | NR | NR | NR | NR | NR | (a) No differences were observed in HAM‐D, HAM‐A, MDQ, HCL‐32, ASI at baseline between groups, and (b) no significant time, group, or time × group differences were observed in HAM‐D and BDI score, or in HAM‐A score, between treatment groups Moderate to severe anxiety was noted in 41.9% (13/32) in the OFS group and 44.8% (14/32) in the tamoxifen group. Both treatment groups had an increase in anxiety levels over time, but there were no significant differences between treatment groups | ||
| 5a/20 | 2a/92 | NR | NR | NR | NR | NR | NR | |
| OFS + chemotherapy vs chemotherapy | ||||||||
| 216a/451^ | 181a/475^ | NR | NR | NR | NR | NR | NR | |
| 8/502^ | 2/494^ | NR | NR | NR | NR | 14c/502 | 6c/494 | |
| NR | NR | NR | NR | NR | NR | 3d/160 | 0d/151 | |
| NR | NR | NR | NR | NR | NR | NR | NR | |
| The median hot flushes score appeared to be worse in the OFS group (goserelin plus chemotherapy) compared to the chemotherapy alone group from approximately 7 months onwards (Figure 1Bernhard 2007). The median hot flushes score seemed to improve from 9 months; however the OFS group still had slightly worse scores than the chemotherapy alone group up until around 30 months (no statistical analyses of treatment comparisons provided in the paper) | NR | NR | NR | NR | NR | NR | ||
| NR | NR | NR | NR | NR | NR | NR | NR | |
| OFS + chemotherapy + tamoxifen vs chemotherapy + tamoxifen | ||||||||
| NR | NR | NR | NR | NR | NR | NR | NR | |
| NR | NR | NR | NR | NR | NR | NR | NR | |
n: number of events; N: number of women studied in each group; NR: not reported.
#Grade 3 or 4 toxicities unless otherwise stated.
*Toxicity data reported on a subset of women in the study (i.e. those from the CRUK enrolment phase).
^Denominator is the randomised number rather than the number of participants who received allocated treatment and assessed while on treatment.
aThis study did not report whether toxicity was assessed using a standardised tool and did not specify the grade of the hot flushes or sweats. ZIPP did not report the grade of vasodilation.
bGrade 3 or 4 neuropsychiatric adverse effects on the NCI CTC; these included outcomes such as anxiety, depression, somnolence, and confusion.
cGrade 3 or 4 neuropsychiatric adverse effects but type of event not specified.
dReported as neuropsychiatric disorders but no further details of how assessed or on severity of the disorders.
In relation to each treatment comparison, ZIPP reported sweating separately for each treatment group as 0% (observation), 1% (tamoxifen), 5% (goserelin), and 5% (goserelin and tamoxifen; Table 1 presents the combined data). ZIPP also reported the incidence of “vasodilation”, which was reported as 0% (observation), 17% (tamoxifen alone), 26% (goserelin alone), and 44% (goserelin and tamoxifen).
Two studies compared OFS and chemotherapy to chemotherapy alone, with any‐grade hot flushes reported in the OFS arm as 48% ‐ Arriagada 2005 ‐ and 79.5% ‐ ECOG 5188, INT‐0101 ‐ and in the control arm as 38% ‐ Arriagada 2005 ‐ and 59.4% ‐ ECOG 5188, INT‐0101. Grade 3 hot flushes were reported in 1.6% in the OFS arm and in 0.4% in the control arm (ECOG 5188, INT‐0101; Table 1).
Four studies reported outcome data for the comparison of OFS and tamoxifen to tamoxifen alone (ABCTCG; E‐3193, INT‐0142; SOFT; ZBCSG Trial B). Two studies reported grade 3 hot flushes in the OFS arm as 13.2% ‐ SOFT ‐ and 16.1% ‐ E‐3193, INT‐0142 ‐ and in the control arm as 7.6% ‐ SOFT ‐ and 4.7% ‐ E‐3193, INT‐0142. One of these studies also reported any‐grade hot flushes as 93.4% in the OFS arm and 79.8% in the control arm (SOFT). ABCTCG reported a higher incidence of night sweats and day sweats in the OFS arm compared to the control arm. Any‐grade sweating was reported in the SOFT study in the OFS arm as 61.8% and in the control arm as 48.3%. No grade 3 or higher events of sweating were reported. ZBCSG Trial B did not state the grade nor any other details for assessing hot flushes; translated information indicated that 25% in the OFS arm and 2.2% in the control arm experienced hot flushes.
Mood
Seven of the 15 studies reported outcomes related to mood, although they were reported in different ways that did not permit a meta‐analysis (refer to Table 1). Two studies reported the incidence of (neuro)psychiatric symptoms. Neuropsychiatric symptoms of grade 3 or higher were reported in the OFS arm as 2.8% ‐ ECOG 5188, INT‐0101 ‐ and 2.3% ‐ E‐3193, INT‐0142 ‐ and in the control arm as 1.2% ‐ ECOG 5188, INT‐0101 ‐ and 2.3% ‐ E‐3193, INT‐0142. ECOG 5188, INT‐0101 also reported any‐grade neuropsychiatric symptoms (including anxiety and depression) in the OFS arm of 32.2% and in the control arm of 16.4%. Psychiatric symptoms of any grade were reported in the OFS arm as 1.9% and in the control arm as 0% (GABG IV‐B‐93). ZIPP reported anxiety, depression, and irritability as a combined measure, with incidence reported as 0% (observation), 2% (tamoxifen), 6% (goserelin), and 6% (goserelin and tamoxifen; see Table 1 for combined data). SOFT reported grade 3 or higher depression in the OFS arm as 4.4% and in the control arm as 3.8%, and the incidence of any‐grade depression in the OFS arm as 51.9% and in the control arm as 46.6%. Yi 2016 reported the incidence of moderate to severe anxiety as 41.9% in the OFS arm and 44.8% in the control arm. ABCTCG reported the OFS arm had a greater incidence of depression and anxiety. See Table 1.
Bone health
Bone health outcomes were reported in one study (SOFT). At a median follow up of 5.6 years, osteoporosis (based on T score < ‐2.5) was reported in the OFS arm as 5.8% and in the control arm as 3.5%. Osteoporosis appears to be worse in the OFS arm than in the control arm (RR 1.16, 95% CI 1.10 to 2.50; 2011 participants; 1 study; low‐certainty evidence). Grade 3 osteoporosis was reported in three participants in the OFS arm and in one participant in the control arm with certainty of evidence considered to be low due to the small number of events and the wide confidence intervals (RR 3.00, 95% CI 0.31 to 28.82; 1 study; 2011 women; Table 1). Any‐grade osteoporosis was reported in the OFS arm as 20%, and in the control arm as 12.3%. In addition, fractures (defined as 'any event') were reported in the OFS arm as 5.4% (54/1005) and 4.9% (49/100 6) at a median follow‐up of 5.6 years (SOFT).
Arthralgia (joint pain)
Arthralgia was specifically reported in ZIPP. Any‐grade arthralgia was reported in 0% (observation), 1% (tamoxifen only), 5% (goserelin only), and 2% (goserelin and tamoxifen; see Table 1 for combined data). SOFT reported on the incidence of musculoskeletal symptoms as any‐grade symptoms in the OFS arm of 75.1% and in the control arm 69.0%, and as grade 3 or higher in the OFS arm of 5.5% and in the control arm 6.3%.
Sexual function
Four of the 15 studies reported outcomes related to sexual function. Any‐grade vaginal dryness was reported in SOFT in the OFS arm as 49.8% and in the control arm as 41.8%, with no grade 3 or higher events. Vaginal dryness was reported in E‐3193, INT‐0142 as grade 3 or higher toxicity in 0.6% (1 of 174 women) of the OFS arm, and no events were reported in the control arm. This was reported descriptively in ZIPP as greatest in the goserelin arm over time compared to each other arm, with the tamoxifen only arm and the goserelin and tamoxifen arm reporting more vaginal dryness than the observation arm. Any‐grade decreases in libido were reported in SOFT in the OFS arm as 47.5% and in the control arm as 42.4%. In ABCTCG, the OFS arm was reported to have more vaginal dryness than the control arm, although there were no differences in sexual function. No numerical data were provided in ABCTCG. See Table 2.
| Study | Sexual function/vaginal dryness | Cardiovascular risk, DVT/PE | Cognitive function | Treatment‐related death | Other toxicities | ||||
| OFS (n/N) | Comparator (n/N) | OFS (n/N) | Comparator (n/N) | OFS (n/N) | Comparator (n/N) | OFS (n/N) | Comparator (n/N) | ||
| OFS vs observation | |||||||||
| NR | NR | NR | NR | NR | NR | NR | NR | Weight gain: 55/591 in OFS group, 32/600 in no OFS group | |
| OFS + tamoxifen vs tamoxifen | |||||||||
| Participants in the OFS group (n = 118) experienced more vaginal dryness (P < 0.001) than those in the no OFS group (n = 128) | NR | NR | NR | NR | 0/1063 | 2/1081 | |||
| 1a/174 | 0a/170 | NR | NR | NR | NR | 0/174 | 0/171 | ||
| Decreased libido (any event): 477/1005 Vaginal dryness (any event): 500/1005 | Decreased libido (any event): 427/1006 Vaginal dryness (any event); 421/1006 | Glucose intolerance (Grade 3/4): 14/1005 | Glucose intolerance (Grade 3/4): 3/1006 | NR | NR | NR | NR | Supplementary data in the 2015 trial publication provide incidence of events for a number of toxicities including insomnia, fatigue, and nausea p.27 | |
| NR | NR | NR | NR | NR | NR | NR | NR | In the protocol publication where baseline characteristics were reported, it was reported that "serious adverse events were not observed during the period of intervention or follow‐up" p.585 | |
| NR | NR | NR | NR | NR | NR | NR | NR | ||
| NR | NR | NR | NR | NR | NR | NR | NR | Weight gain noted in 2/20 in the OFS group, 5/92 in the no OFS group | |
| OFS + chemotherapy vs chemotherapy | |||||||||
| NR | NR | NR | NR | NR | NR | NR | NR | "No severe adverse effects were documented" p.395. Body weight was recorded with no difference between OFS and no OFS treatment groups | |
| NR | NR | 14/502 | 16/494 | NR | NR | 1/502 lethal event (cardiomyopathy) during maintenance phase | 6/494 lethal adverse events were recorded during chemotherapy (4 events ‐ 2 sepsis, 1 myocardial infarction, and 1 cardiomyopathy and pneumonia) and in maintenance phase (2 events ‐ suicide, unspecified pulmonary disease) | An increase in weight and hypertension was noted in the OFS group | |
| NR | NR | NR | NR | NR | NR | 0/160 | 0/151 | Leukopenia: 1/160 (OFS), 5/151 (no OFS); emesis/nausea: 1/160 (OFS), 2/151 (no OFS); paravasation: 1/160 (OFS), 0/151 (no OFS). In the OFS group, 2 participants each had the following: wound healing and erysipelas; in the no OFS group, there were 2 participants each with seroma and abscess. One participant in each group had the following: infection, wound pain, endometrial hyperproliferation,, mastopathy, thrombophlebitis, hyponatraemia, stomatitis, vertigo, infection, and fever | |
| NR | NR | NR | NR | NR | NR | NR | NR | ||
| NR | NR | NR | NR | NR | NR | 0/357^ | 0/360^ | In the OFS group: 1/360 life‐threatening (suicidal) depression reported after 6 months of chemotherapy and 4 goserelin implants. Alopecia reported but only in those participants who had chemotherapy; information was not presented separately for each group. Weight gain was mentioned only in the OFS group and further details were provided in the trial publication | |
| NR | NR | NR | NR | NR | NR | 0/148^ | 0/140^ | Leukopenia: 27/148 (OFS), 22/140 (no OFS); neuropathy: 4/148 (OFS), 9/140 (no OFS); mucositis: 20/148 (OFS), 19/148 (no OFS); fatigue: 29/148 (OFS), 29/140 (no OFS) | |
| OFS + chemotherapy + tamoxifen vs chemotherapy + tamoxifen | |||||||||
| NR | NR | NR | NR | NR | NR | NR | NR | ||
| NR | NR | NR | NR | NR | NR | NR | NR | ||
n: number of events; N: number of women studied in each group; NR: not reported.
aGrade 3 or 4 vaginal dryness reported using the NCI CTC.
^Denominator is the randomised number rather than the number of participants who received the allocated treatment and assessed while on treatment.
Cardiovascular risk/DVT/PE
Two of the 15 studies reported outcomes related to cardiovascular conditions. Any‐grade glucose intolerance was reported in the OFS arm as 44.2% ‐ ECOG 5188, INT‐0101 ‐ and 3.5% ‐ SOFT ‐ and in the control arm as 36.9% ‐ ECOG 5188, INT‐0101 ‐ and 1.8% ‐ SOFT. Grade 3 or higher glucose intolerance was reported in the OFS arm as 2.8% ‐ ECOG 5188, INT‐0101 ‐ and 1.4% ‐ SOFT ‐ and in the control arm as 3.2% ‐ ECOG 5188, INT‐0101 ‐ and 0.3% ‐ SOFT. Any‐grade hypertension was reported in the OFS arm as 22.2% and in the control arm as 17.2% (SOFT). Grade 3 or higher hypertension was reported in the OFS arm as 7.5% and in the control arm as 5.4% (SOFT). Thrombosis or embolism of any grade was reported in the OFS arm as 2.3% and in the control arm as 2.2% (SOFT). See Table 2.
Cognitive function
Two of 15 studies reported outcomes related to cognitive function. One study within the ZIPP collaboration reported that memory and concentration problems were not affected by either treatment. SOFT reported no difference in the objective measurement of cognitive function between arms, although a decline in self‐reported cognitive function was noted in the OFS arm compared to the control arm. See Table 2.
Treatment‐related death
Six of 15 studies reported outcomes for treatment‐related deaths (ABCTCG; ECOG 5188, INT‐0101; E‐3193, INT‐0142; GABG IV‐B‐93; IBCSG VIII; SWOG 1996). ABCTCG reported two deaths due to chemotherapy toxicity and no treatment‐related deaths in the OFS treatment group. ECOG 5188, INT‐0101 reported four lethal adverse events during chemotherapy from sepsis (two participants), myocardial infarction (one participant), and cardiomyopathy and pneumonitis (one participant). Four lethal adverse events were reported during the maintenance phase ‐ one in the OFS arm (cardiomyopathy) and two in the control arm (suicide and unspecified pulmonary disease, respectively). E‐3193, INT‐0142, GABG IV‐B‐93, IBCSG VIII, and SWOG 1996 reported no treatment‐related deaths. See Table 2.
Other toxicities
Seven of the 15 studies also reported other toxicities not otherwise mentioned above (see Table 2). GABG IV‐B‐93, IBCSG II, and SWOG 1996 reported on toxicities that appeared to be chemotherapy related, including myelosuppression, gastrointestinal symptoms, peripheral neuropathy, rash, and lung problems. ZIPP reported weight gain as 0% (observation), 7% (tamoxifen), 4% (goserelin), and 11% (goserelin and tamoxifen). Goserelin and goserelin and tamoxifen arms had a higher incidence of problems with body image than the tamoxifen alone arm. There was no difference in sleep or fatigue between arms. ZBCSG Trial B reported weight gain of 10% in the OFS group and 5.4% in the control group, and so too did IBCSG VIII; Arriagada 2005 recorded no differences between treatment groups in relation to weight gain. ECOG 5188, INT‐0101 reported an increase in hypertension and weight in the OFS group. In SOFT, there were higher rates of any‐grade insomnia in the OFS arm (57.2%) than in the control arm (46.3%) and of grade 3 or higher events of insomnia (4.6% versus 2.9%, respectively).
Quality of life
Four out of 15 studies collected data on quality of life using validated tools (ABCTCG; E‐3193, INT‐0142; IBCSG VIII; SOFT), and one study collected quality of life‐type information without describing a validated tool (ZBCSG Trial B). Two studies reported worse quality of life indicators (i.e. vaginal dryness, day and night sweats) in the OFS group than in the no OFS group (ABCTCG; E‐3193, INT‐0142). The other two studies indicated worsening of symptoms (e.g. vasomotor, gynaecological, vaginal dryness, decline in sexual interest, bone and joint pain, and weight gain); however these side effects were reported in both OFS and no OFS groups (IBCSG VIII; SOFT). The study that did not use a validated quality of life tool described no considerable differences between groups (ZBCSG Trial B).
Compliance with treatment
In this review, compliance with treatment referred to participants who received their allocated treatment and did not stop treatment early due to toxicities. Six of 15 studies reported the number of women who received LHRH agonists and discontinued due to toxicity (Arriagada 2005; E‐3193, INT‐0142; ECOG 5188, INT‐0101; GABG IV‐B‐93; IBCSG VIII; SOFT). Over 78% of women completed the intended course of OFS in these studies: Arriagada 2005: 85% (151/177), E‐3193, INT‐0142: 90% (153/170), GABG IV‐B‐93: 92.5% (296/320), ECOG 5188, INT‐0101: 84.3% (348/413), IBCSG VIII: 93% (332/357), and SOFT: 78% (792/1015). Two of the six studies provided hormonal treatment in the control arm, with treatment completion rates of 93.8% (155/167) ‐ E‐3193, INT‐0142 ‐and 77.5% (789/1018) ‐ SOFT.
Discusión
Resumen de los resultados principales
En las mujeres premenopáusicas con cáncer de mama temprano positivo para los receptores hormonales, el agregado de la supresión de la función ovárica (SFO) dio lugar a una modesta mejora de la supervivencia general y de la supervivencia libre de enfermedad, así como a una reducción general de los cánceres de mama contralaterales. Este efecto beneficioso observado sobre la supervivencia general y la supervivencia libre de enfermedad persistió entre las participantes que se asignaron al azar a SFO, SFO y tratamiento endocrino, y SFO combinado con quimioterapia y tratamiento endocrino. Los regímenes de quimioterapia incluyeron ciclofosfamida, metotrexato y fluorouracilo (CMF) y regímenes de antraciclina, y el tratamiento endocrino en los estudios incluidos fue el tamoxifeno. Ningún estudio examinó la SFO más inhibidores de la aromatasa versus ninguna SFO más inhibidores de la aromatasa, como era de esperar en las mujeres premenopáusicas. En algunos subgrupos, el agregado de SFO dio lugar a una mejora en la supervivencia libre de enfermedad, sin mejorar la supervivencia general. Se incluyeron participantes menores de 40 años, participantes que tenían menos de tres años de SFO y participantes que no recibieron quimioterapia. Los resultados se pueden deber a que el número de participantes en estos subgrupos es menor, en especial porque no todos los estudios estratificaron sus resultados por estos factores, y porque la duración del seguimiento fue más corta en estos grupos.
La SFO probablemente dio lugar a una mayor incidencia de sofocos de cualquier grado y graves y puede aumentar el riesgo de osteoporosis, aunque puede ser necesario un seguimiento más largo para evaluar la salud ósea. En la mayoría de los estudios se observaron diferencias mínimas en cuanto a la artralgia, el estado de ánimo, la función sexual, los resultados cardiovasculares y la función cognitiva, aunque pocos estudios informaron de manera adecuada los datos de toxicidad. El agregado de SFO no pareció aumentar las muertes relacionadas con el tratamiento, ni las segundas neoplasias malignas.
Compleción y aplicabilidad general de las pruebas
La generalización de estos resultados puede estar afectada por la administración de regímenes de quimioterapia no desactualizados, ya que algunos de los estudios incluidos fueron antiguos, y las mujeres con cánceres positivos a receptores hormonales o HER2 pueden no haber recibido tratamiento hormonal ni tratamiento dirigido a HER2 como parte de la atención estándar. Entre los estudios que informaron sobre el régimen de quimioterapia administrado, se calculó que el 72% (3091/4296) de las mujeres recibieron una antraciclina; en un caso, las antraciclinas se administraron solo en las mujeres de alto riesgo (con cuatro a nueve ganglios linfáticos afectados). La inclusión del estudio SOFT como un estudio más reciente en el que se evaluó el receptor HER2 apoya el agregado de SFO en el contexto de los paradigmas de tratamiento modernos.
La mayoría de los estudios informaron sobre los resultados supervivencia general y supervivencia libre de enfermedad. La mayoría de los estudios no informaron sobre la incidencia de cáncer de mama contralateral o de una segunda neoplasia maligna. La mayoría de los estudios no informaron sobre resultados de toxicidad. Hubo variaciones en cuanto a los efectos secundarios que se recogieron y la forma en que se informaron. Las toxicidades se informaron como incidencia de toxicidad de cualquier grado, incidencia de toxicidad de alto grado o una descripción de las toxicidades, sin proporcionar una comparación numérica. No se obtuvo la intensidad de los efectos secundarios en el tiempo. Es de señalar que solo un estudio informó sobre los resultados de la salud ósea, definidos por los umbrales de densidad mineral ósea. Aunque también se informó sobre la incidencia de fracturas, no se registraron las tasas de administración de bifosfonatos. Es más probable que la salud ósea esté afectada a largo plazo, en lugar de a corto o mediano plazo; por lo tanto, se requiere un seguimiento de mayor duración. La mayoría de los estudios no midieron resultados de la calidad de vida. Cuando se midió la calidad de vida, cada uno de los cinco estudios utilizó un instrumento diferente o no informó si se había utilizado un cuestionario validado.
Calidad de la evidencia
Esta revisión sistemática proporciona evidencia de 15 estudios que incluyeron 11 538 mujeres premenopáusicas con cáncer de mama temprano positivo para los receptores hormonales. Evidencia de alta certeza respalda el agregado de la SFO al tratamiento estándar en el entorno adyuvante, con un aumento de la supervivencia general y de la supervivencia libre de enfermedad. Fueron estudios abiertos, lo que puede aumentar el riesgo de sesgo, en particular para las evaluaciones de la toxicidad y la calidad de vida. Hubo heterogeneidad clínica evidente para la duración de la SFO, el método de SFO, la positividad ganglionar de los cánceres y el tipo de quimioterapia administrada (cuando fue aplicable).
Sesgos potenciales en el proceso de revisión
No se incluyeron los estudios que administraron la SFO durante 12 meses o menos e informaron sobre variables principales de evaluación centradas en la fertilidad, ya que estuvieron fuera del alcance de esta revisión. Pueden haber proporcionado evidencia que indique la presencia de una relación dependiente de la dosis entre el agregado de la SFO y los resultados de eficacia.
Cuando la población total de los estudios incluyó cánceres positivos y negativos para los receptores hormonales, y cuando los resultados no se estratificaron por el estado de las hormonas, se utilizaron los datos de la población total (siempre y cuando la incidencia de los cánceres positivos para los receptores hormonales fuera superior al 50%). Lo anterior puede haber subestimado la eficacia del agregado de la SFO, que se considera que solo afecta los resultados en los cánceres positivos para los receptores hormonales.
Acuerdos y desacuerdos con otros estudios o revisiones
Los presentes resultados son consistentes con el metanálisis del EBCTCG (EBCTCG 2005), en el que se examinaron los datos de participantes individuales de 8000 mujeres y se determinó que la SFO dio lugar a una reducción de la recidiva y la mortalidad por cáncer de mama, aunque el efecto fue menor cuando la SFO se administró junto con otros tratamientos. La población incluyó el 47% de las participantes con estado hormonal no comprobado que se asignaron al azar a un estudio clínico antes del año 2000, en estudios que comenzaron antes de 1995. La presente revisión incluye mujeres que tenían cánceres positivo para los receptores hormonales y que habían recibido tratamientos más modernos como las antraciclinas. Algunos de los estudios incluidos en la revisión de los datos de las participantes individuales no se incluyeron en esta revisión Cochrane porque no se estratificaron según el estado de los receptores hormonales, o no proporcionaron detalles suficientes acerca de la duración de la SFO. En esta revisión Cochrane se incluyeron estudios adicionales en el metanálisis que reflejaron los datos agregados y los nuevos estudios publicados desde 2005.
Otras revisiones que examinaron la función de la SFO en el tratamiento adyuvante del cáncer de mama temprano incluyen Cheer 2005 y Chlebowski 2017.
Cheer 2005 examinó los efectos de la goserelina en el tratamiento del cáncer de mama temprano, e incluyó tres de los estudios que se habían identificado aquí (ECOG 5188, INT‐0101; IBCSG VIII; ZIPP). La revisión incluye un seguimiento más largo para dos de estos estudios y examina la función de la SFO en lugar de la goserelina sola. Los cuatro estudios restantes incluidos en la revisión Cheer 2005 no cumplieron los criterios de elegibilidad e incluyeron comparaciones de goserelina con quimioterapia, o agregaron tamoxifeno y goserelina al brazo de intervención.
Chlebowski 2017 examinó los efectos de la SFO en combinación con el tratamiento adyuvante endocrina combinada e incluyó dos de los estudios identificados en la búsqueda actual (E‐3193, INT‐0142; SOFT). La presente revisión incluye un seguimiento más largo para SOFT y examina el efecto más amplio de la SFO considerando el agregado de la SFO a la observación, la quimioterapia o el tamoxifeno, en lugar de solo a otros agentes hormonales. Los dos estudios restantes incluidos en la revisión de Chlebowski 2017 no cumplieron los criterios de elegibilidad e incluyeron la comparación de SFO e inhibidores de la aromatasa con SFO y tamoxifeno.
Study flow diagram: 2019 review update.
Risk of bias summary: review authors' judgements about each risk of bias item for each included study.
Forest plot of comparison: 1 OFS versus no OFS, outcome: 1.1 Overall survival.
Funnel plot of comparison: 1 OFS versus no OFS, outcome: 1.1 Overall survival.
Forest plot of comparison: 1 OFS versus no OFS, outcome: 1.2 Disease‐free survival.
Funnel plot of comparison: 1 OFS versus no OFS, outcome: 1.2 Disease‐free survival.
Forest plot of comparison: 1 OFS versus no OFS, outcome: 1.5 Hot flushes.
Comparison 1 OFS versus no OFS, Outcome 1 Overall survival.
Comparison 1 OFS versus no OFS, Outcome 2 Disease‐free survival.
Comparison 1 OFS versus no OFS, Outcome 3 Contralateral breast cancer.
Comparison 1 OFS versus no OFS, Outcome 4 Second malignancy.
Comparison 1 OFS versus no OFS, Outcome 5 Hot flushes.
Comparison 2 Duration of OFS: < 3 years vs ≥ 3 years, Outcome 1 Overall survival.
Comparison 2 Duration of OFS: < 3 years vs ≥ 3 years, Outcome 2 Disease‐free survival.
Comparison 3 Age of studied population: < 40 years vs ≥ 40 years, Outcome 1 Overall survival.
Comparison 3 Age of studied population: < 40 years vs ≥ 40 years, Outcome 2 Disease‐free survival.
Comparison 4 Chemotherapy use: yes or no, Outcome 1 Overall survival.
Comparison 4 Chemotherapy use: yes or no, Outcome 2 Disease‐free survival.
Comparison 5 Method of OFS, Outcome 1 Overall survival.
Comparison 5 Method of OFS, Outcome 2 Disease‐free survival.
Comparison 6 Lymph node status: positive or negative, Outcome 1 Overall survival.
Comparison 6 Lymph node status: positive or negative, Outcome 2 Disease‐free survival.
| OFS compared to no OFS for adjuvant treatment of early breast cancer | ||||||
| Patient or population: women with early breast cancer | ||||||
| Outcomes | Anticipated absolute effects* (95% CI) | Relative effect | № of participants | Certainty of the evidence | Comments | |
| Risk with no OFS | Risk with OFS | |||||
| Overall survival (OS) Median follow‐up: range 5.3 to 12.1 years (*baseline risk at 5 and 10 years estimated from the control arm in 8 and 7 studies, respectively) | 5‐year risk of death* | HR 0.86 | 10,374 | ⊕⊕⊕⊕ | ||
| 110 per 1000 | 95 per 1000 | |||||
| 10‐year risk of death* | ||||||
| 310 per 1000 | 273 per 1000 | |||||
| Disease‐free survival (DFS) (*baseline risk at 5 and 10 years estimated from the control arm in 9 and 7 studies, respectively) | 5‐year risk of recurrence* | HR 0.83 | 8899 | ⊕⊕⊕⊕ | ||
| 250 per 1000 | 212 per 1000 | |||||
| 10‐year risk of recurrence* | ||||||
| 440 per 1000 | 382 per 1000 | |||||
| Toxicity ‐ hot flushes (a combination of "grade 3/4" and "any grade" toxicity) Follow‐up: range 2 to 5 years | 97 per 1000 | 154 per 1000 | RR 1.60 | 5581 | ⊕⊕⊕⊝ | An additional 2 studies (studied population: 246 in ABCTCG, not detailed in IBCSG VIII) reported a higher number of hot flushes in the OFS group than in the no OFS group. These results are consistent with the overall effect estimate |
| Toxicity ‐ mood Follow‐up: range 1 year to 9.6 years | 7 out of 15 studies reported Grade 3 or 4 depression, anxiety, or neuropsychiatric symptoms or did not report the grade of these side effects. Two studies ‐ ABCTCG; E‐3193, INT‐0142 ‐ reported an increase in depression, anxiety, and neuropsychiatric symptoms in the OFS group compared to the non‐OFS group. Five studies ‐ ECOG 5188, INT‐0101; GABG IV‐B‐93; SOFT; Yi 2016; ZIPP ‐ indicated an increase in anxiety in both treatment groups (but no differences between groups) or no difference overall in symptoms over time or between treatment groups | ‐ | 5354 | ⊕⊕⊕⊝ | ||
| Toxicity ‐ bone health (osteoporosis defined by a T score < ‐2.5) Follow‐up: median 5.6 years | 35 per 1000 | 58 per 1000 | RR 1.66 (1.10 to 2.50) | 2011 | ⊕⊕⊝⊝c | |
| Contralateral breast cancer Follow‐up: 4.75 to median 12.1 years | 31 per 1000 | 23 per 1000 | RR 0.75 | 7856 | ⊕⊕⊕⊝ | |
| Quality of life | Four out of 15 studies collected data on quality of life using validated tools (ABCTCG; E‐3193, INT‐0142; IBCSG VIII; SOFT), and 1 study collected quality of life‐type information without describing a validated tool (ZBCSG Trial B). Two studies ‐ ABCTCG; E‐3193, INT‐0142 ‐ reported worse quality of life indicators (i.e. vaginal dryness, day and night sweats) in the OFS group than in the no OFS group. The other 2 studies ‐ IBCSG VIII; SOFT ‐ indicated worsening of symptoms (e.g. vasomotor, gynaecological, vaginal dryness, decline in sexual interest, bone and joint pain, weight gain); however these side effects were reported in both OFS and no OFS groups. The study that did not use a validated quality of life tool described no considerable differences between groups | ‐ | Estimated to be 2996 (5 RCTs) | ⊕⊕⊝⊝ | ||
| *The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). | ||||||
| GRADE Working Group grades of evidence. | ||||||
| aThree studies reported Grade 3 or 4 hot flushes using a standardised toxicity symptom scale (E‐3193, INT‐0142; ECOG 5188, INT‐0101; SOFT), and the other three studies that contributed to the meta‐analysis reported any grade of hot flushes/sweats without reporting the scale used (Arriagada 2005; ZBCSG Trial B; ZIPP). This outcome was downgraded due to variability in reporting of hot flushes across studies and uncertainty as to whether unblinding of treatment allocation may have affected patient reporting and concerns about selective outcome reporting. Therefore downgraded by two points overall. | ||||||
| Study | Hot flushes/menopausal symptoms# | Bone health: bone density/fractures | Arthralgias (joint pain) | Mood (anxiety, depression, other as indicated in footnotes) | ||||
| OFS (n/N) | Comparator (n/N) | OFS (n/N) | Comparator (n/N) | OFS (n/N) | Comparator (n/N) | OFS (n/N) | Comparator (n/N) | |
| OFS vs observation | ||||||||
| ZIPP* | Reported as: | Reported as: | NR | NR | 17/591 | 6/600 | 34/591 | 11/600 Reported as anxiety/depression/irritability |
| OFS + tamoxifen vs tamoxifen | ||||||||
| Participants in the OFS group (n = 118) experienced more menopausal symptoms (night sweats, P = 0.005) and day sweats (P < 0.001) than those in the no OFS group (n = 128) | NR | NR | NR | NR | Participants in the OFS group had increased depression (P = 0.05) and anxiety (P = 0.04) over 30 months compared to those in the no OFS group (total number of participants = 436, data not reported separately by treatment group) | |||
| 28/174 | 8/171 | NR | NR | NR | NR | 4b/174 | 4b/171 | |
| 133/1005 | 76/1006 | 3/1005 | 1/1006 | 55/1005 | 63/1006 | 44/1005 | 38/1006 | |
| NR | NR | NR | NR | NR | NR | NR | NR | |
| NR | NR | NR | NR | NR | NR | (a) No differences were observed in HAM‐D, HAM‐A, MDQ, HCL‐32, ASI at baseline between groups, and (b) no significant time, group, or time × group differences were observed in HAM‐D and BDI score, or in HAM‐A score, between treatment groups Moderate to severe anxiety was noted in 41.9% (13/32) in the OFS group and 44.8% (14/32) in the tamoxifen group. Both treatment groups had an increase in anxiety levels over time, but there were no significant differences between treatment groups | ||
| 5a/20 | 2a/92 | NR | NR | NR | NR | NR | NR | |
| OFS + chemotherapy vs chemotherapy | ||||||||
| 216a/451^ | 181a/475^ | NR | NR | NR | NR | NR | NR | |
| 8/502^ | 2/494^ | NR | NR | NR | NR | 14c/502 | 6c/494 | |
| NR | NR | NR | NR | NR | NR | 3d/160 | 0d/151 | |
| NR | NR | NR | NR | NR | NR | NR | NR | |
| The median hot flushes score appeared to be worse in the OFS group (goserelin plus chemotherapy) compared to the chemotherapy alone group from approximately 7 months onwards (Figure 1Bernhard 2007). The median hot flushes score seemed to improve from 9 months; however the OFS group still had slightly worse scores than the chemotherapy alone group up until around 30 months (no statistical analyses of treatment comparisons provided in the paper) | NR | NR | NR | NR | NR | NR | ||
| NR | NR | NR | NR | NR | NR | NR | NR | |
| OFS + chemotherapy + tamoxifen vs chemotherapy + tamoxifen | ||||||||
| NR | NR | NR | NR | NR | NR | NR | NR | |
| NR | NR | NR | NR | NR | NR | NR | NR | |
| n: number of events; N: number of women studied in each group; NR: not reported. #Grade 3 or 4 toxicities unless otherwise stated. | ||||||||
| Study | Sexual function/vaginal dryness | Cardiovascular risk, DVT/PE | Cognitive function | Treatment‐related death | Other toxicities | ||||
| OFS (n/N) | Comparator (n/N) | OFS (n/N) | Comparator (n/N) | OFS (n/N) | Comparator (n/N) | OFS (n/N) | Comparator (n/N) | ||
| OFS vs observation | |||||||||
| NR | NR | NR | NR | NR | NR | NR | NR | Weight gain: 55/591 in OFS group, 32/600 in no OFS group | |
| OFS + tamoxifen vs tamoxifen | |||||||||
| Participants in the OFS group (n = 118) experienced more vaginal dryness (P < 0.001) than those in the no OFS group (n = 128) | NR | NR | NR | NR | 0/1063 | 2/1081 | |||
| 1a/174 | 0a/170 | NR | NR | NR | NR | 0/174 | 0/171 | ||
| Decreased libido (any event): 477/1005 Vaginal dryness (any event): 500/1005 | Decreased libido (any event): 427/1006 Vaginal dryness (any event); 421/1006 | Glucose intolerance (Grade 3/4): 14/1005 | Glucose intolerance (Grade 3/4): 3/1006 | NR | NR | NR | NR | Supplementary data in the 2015 trial publication provide incidence of events for a number of toxicities including insomnia, fatigue, and nausea p.27 | |
| NR | NR | NR | NR | NR | NR | NR | NR | In the protocol publication where baseline characteristics were reported, it was reported that "serious adverse events were not observed during the period of intervention or follow‐up" p.585 | |
| NR | NR | NR | NR | NR | NR | NR | NR | ||
| NR | NR | NR | NR | NR | NR | NR | NR | Weight gain noted in 2/20 in the OFS group, 5/92 in the no OFS group | |
| OFS + chemotherapy vs chemotherapy | |||||||||
| NR | NR | NR | NR | NR | NR | NR | NR | "No severe adverse effects were documented" p.395. Body weight was recorded with no difference between OFS and no OFS treatment groups | |
| NR | NR | 14/502 | 16/494 | NR | NR | 1/502 lethal event (cardiomyopathy) during maintenance phase | 6/494 lethal adverse events were recorded during chemotherapy (4 events ‐ 2 sepsis, 1 myocardial infarction, and 1 cardiomyopathy and pneumonia) and in maintenance phase (2 events ‐ suicide, unspecified pulmonary disease) | An increase in weight and hypertension was noted in the OFS group | |
| NR | NR | NR | NR | NR | NR | 0/160 | 0/151 | Leukopenia: 1/160 (OFS), 5/151 (no OFS); emesis/nausea: 1/160 (OFS), 2/151 (no OFS); paravasation: 1/160 (OFS), 0/151 (no OFS). In the OFS group, 2 participants each had the following: wound healing and erysipelas; in the no OFS group, there were 2 participants each with seroma and abscess. One participant in each group had the following: infection, wound pain, endometrial hyperproliferation,, mastopathy, thrombophlebitis, hyponatraemia, stomatitis, vertigo, infection, and fever | |
| NR | NR | NR | NR | NR | NR | NR | NR | ||
| NR | NR | NR | NR | NR | NR | 0/357^ | 0/360^ | In the OFS group: 1/360 life‐threatening (suicidal) depression reported after 6 months of chemotherapy and 4 goserelin implants. Alopecia reported but only in those participants who had chemotherapy; information was not presented separately for each group. Weight gain was mentioned only in the OFS group and further details were provided in the trial publication | |
| NR | NR | NR | NR | NR | NR | 0/148^ | 0/140^ | Leukopenia: 27/148 (OFS), 22/140 (no OFS); neuropathy: 4/148 (OFS), 9/140 (no OFS); mucositis: 20/148 (OFS), 19/148 (no OFS); fatigue: 29/148 (OFS), 29/140 (no OFS) | |
| OFS + chemotherapy + tamoxifen vs chemotherapy + tamoxifen | |||||||||
| NR | NR | NR | NR | NR | NR | NR | NR | ||
| NR | NR | NR | NR | NR | NR | NR | NR | ||
| n: number of events; N: number of women studied in each group; NR: not reported. aGrade 3 or 4 vaginal dryness reported using the NCI CTC. | |||||||||
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 1 Overall survival Show forest plot | 11 | 10374 | Hazard Ratio (Fixed, 95% CI) | 0.86 [0.78, 0.94] |
| 1.1 OFS vs observation | 1 | 2706 | Hazard Ratio (Fixed, 95% CI) | 0.83 [0.71, 0.97] |
| 1.2 OFS + chemotherapy vs chemotherapy | 5 | 3087 | Hazard Ratio (Fixed, 95% CI) | 0.95 [0.82, 1.09] |
| 1.3 OFS + tamoxifen vs tamoxifen | 4 | 3299 | Hazard Ratio (Fixed, 95% CI) | 0.74 [0.59, 0.93] |
| 1.4 OFS + chemotherapy + tamoxifen vs chemotherapy + tamoxifen | 1 | 1282 | Hazard Ratio (Fixed, 95% CI) | 0.31 [0.10, 0.94] |
| 2 Disease‐free survival Show forest plot | 10 | 8899 | Hazard Ratio (Fixed, 95% CI) | 0.83 [0.77, 0.90] |
| 2.1 OFS vs observation | 1 | 2706 | Hazard Ratio (Fixed, 95% CI) | 0.82 [0.73, 0.92] |
| 2.2 OFS + chemotherapy vs chemotherapy | 5 | 2450 | Hazard Ratio (Fixed, 95% CI) | 0.90 [0.79, 1.01] |
| 2.3 OFS + tamoxifen vs tamoxifen | 3 | 2461 | Hazard Ratio (Fixed, 95% CI) | 0.76 [0.63, 0.92] |
| 2.4 OFS + chemotherapy + tamoxifen vs chemotherapy + tamoxifen | 1 | 1282 | Hazard Ratio (Fixed, 95% CI) | 0.69 [0.48, 0.99] |
| 3 Contralateral breast cancer Show forest plot | 9 | 9138 | Risk Ratio (Fixed, 95% CI) | 0.75 [0.58, 0.98] |
| 3.1 OFS vs observation | 1 | 2710 | Risk Ratio (Fixed, 95% CI) | 0.75 [0.47, 1.17] |
| 3.2 OFS + chemotherapy vs chemotherapy | 5 | 3012 | Risk Ratio (Fixed, 95% CI) | 0.82 [0.56, 1.20] |
| 3.3 OFS + tamoxifen vs tamoxifen | 1 | 2033 | Risk Ratio (Fixed, 95% CI) | 0.56 [0.29, 1.07] |
| 3.4 OFS + chemotherapy + tamoxifen vs chemotherapy + tamoxifen | 2 | 1383 | Risk Ratio (Fixed, 95% CI) | 1.02 [0.26, 4.06] |
| 4 Second malignancy Show forest plot | 7 | 6327 | Risk Ratio (Fixed, 95% CI) | 0.89 [0.64, 1.25] |
| 4.1 OFS + chemotherapy vs chemotherapy | 5 | 3012 | Risk Ratio (Fixed, 95% CI) | 0.94 [0.60, 1.45] |
| 4.2 OFS + tamoxifen vs tamoxifen | 1 | 2033 | Risk Ratio (Fixed, 95% CI) | 0.96 [0.55, 1.67] |
| 4.3 OFS + chemotherapy + tamoxifen vs chemotherapy + tamoxifen | 1 | 1282 | Risk Ratio (Fixed, 95% CI) | 0.38 [0.10, 1.43] |
| 5 Hot flushes Show forest plot | 6 | 5581 | Risk Ratio (M‐H, Fixed, 95% CI) | 1.60 [1.41, 1.82] |
| 5.1 OFS vs observation | 1 | 1191 | Risk Ratio (M‐H, Fixed, 95% CI) | 6.09 [2.38, 15.59] |
| 5.2 OFS + chemotherapy vs chemotherapy | 2 | 1922 | Risk Ratio (M‐H, Fixed, 95% CI) | 1.29 [1.11, 1.49] |
| 5.3 OFS + tamoxifen vs tamoxifen | 3 | 2468 | Risk Ratio (M‐H, Fixed, 95% CI) | 1.99 [1.56, 2.55] |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 1 Overall survival Show forest plot | 10 | 9536 | Hazard Ratio (Fixed, 95% CI) | 0.86 [0.78, 0.95] |
| 1.1 < 3 years of OFS | 5 | 4956 | Hazard Ratio (Fixed, 95% CI) | 0.79 [0.69, 0.91] |
| 1.2 ≥ 3 years of OFS | 5 | 4580 | Hazard Ratio (Fixed, 95% CI) | 0.93 [0.81, 1.07] |
| 2 Disease‐free survival Show forest plot | 10 | 8899 | Hazard Ratio (Fixed, 95% CI) | 0.83 [0.77, 0.90] |
| 2.1 < 3 years of OFS | 5 | 4956 | Hazard Ratio (Fixed, 95% CI) | 0.80 [0.72, 0.88] |
| 2.2 ≥ 3 years of OFS | 5 | 3943 | Hazard Ratio (Fixed, 95% CI) | 0.88 [0.78, 0.98] |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 1 Overall survival Show forest plot | 2 | Hazard Ratio (Fixed, 95% CI) | Subtotals only | |
| 1.1 < 40 years | 2 | 394 | Hazard Ratio (Fixed, 95% CI) | 0.73 [0.51, 1.04] |
| 1.2 ≥ 40 years | 2 | 1175 | Hazard Ratio (Fixed, 95% CI) | 0.89 [0.69, 1.14] |
| 2 Disease‐free survival Show forest plot | 3 | Hazard Ratio (Fixed, 95% CI) | Subtotals only | |
| 2.1 < 40 years | 3 | 1764 | Hazard Ratio (Fixed, 95% CI) | 0.65 [0.50, 0.83] |
| 2.2 ≥ 40 years | 3 | 1504 | Hazard Ratio (Fixed, 95% CI) | 0.95 [0.78, 1.15] |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 1 Overall survival Show forest plot | 9 | 6739 | Hazard Ratio (Fixed, 95% CI) | 0.87 [0.77, 0.97] |
| 1.1 Chemotherapy | 8 | 5453 | Hazard Ratio (Fixed, 95% CI) | 0.86 [0.76, 0.97] |
| 1.2 No chemotherapy | 3 | 1286 | Hazard Ratio (Fixed, 95% CI) | 0.89 [0.62, 1.28] |
| 2 Disease‐free survival Show forest plot | 9 | 6102 | Hazard Ratio (Fixed, 95% CI) | 0.83 [0.76, 0.90] |
| 2.1 Chemotherapy | 8 | 4816 | Hazard Ratio (Fixed, 95% CI) | 0.85 [0.77, 0.93] |
| 2.2 No chemotherapy | 3 | 1286 | Hazard Ratio (Fixed, 95% CI) | 0.73 [0.59, 0.90] |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 1 Overall survival Show forest plot | 9 | Hazard Ratio (Fixed, 95% CI) | Subtotals only | |
| 1.1 OFS via surgery vs no OFS | 2 | 415 | Hazard Ratio (Fixed, 95% CI) | 0.86 [0.57, 1.28] |
| 1.2 OFS via LHRH agonists vs no OFS | 8 | 8101 | Hazard Ratio (Fixed, 95% CI) | 0.80 [0.71, 0.89] |
| 1.3 OFS via RT vs no OFS | 1 | 77 | Hazard Ratio (Fixed, 95% CI) | 1.75 [0.50, 6.16] |
| 2 Disease‐free survival Show forest plot | 9 | Hazard Ratio (Fixed, 95% CI) | Subtotals only | |
| 2.1 OFS via surgery vs no OFS | 2 | 415 | Hazard Ratio (Fixed, 95% CI) | 0.96 [0.70, 1.30] |
| 2.2 OFS via LHRH agonists vs no OFS | 8 | 8101 | Hazard Ratio (Fixed, 95% CI) | 0.81 [0.75, 0.88] |
| 2.3 OFS via RT vs no OFS | 1 | 77 | Hazard Ratio (Fixed, 95% CI) | 0.94 [0.28, 3.13] |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 1 Overall survival Show forest plot | 10 | 10283 | Hazard Ratio (Fixed, 95% CI) | 0.86 [0.78, 0.94] |
| 1.1 Positive | 7 | 7340 | Hazard Ratio (Fixed, 95% CI) | 0.89 [0.81, 0.99] |
| 1.2 Negative | 3 | 2943 | Hazard Ratio (Fixed, 95% CI) | 0.69 [0.53, 0.88] |
| 2 Disease‐free survival Show forest plot | 9 | 8808 | Hazard Ratio (Fixed, 95% CI) | 0.83 [0.77, 0.90] |
| 2.1 Positive | 6 | 5865 | Hazard Ratio (Fixed, 95% CI) | 0.86 [0.79, 0.93] |
| 2.2 Negative | 3 | 2943 | Hazard Ratio (Fixed, 95% CI) | 0.75 [0.64, 0.89] |
| Study | Questionnaire used | Summary of findings | Follow‐up period |
| OFS versus observation | |||
| NR | NR | NR | |
| OFS + tamoxifen vs tamoxifen | |||
| EORTC QLQ‐C30, BR23 breast cancer module | 246 out of 2144 participants agreed to take part in the Quality of Life substudy. A narrative synthesis of the key findings was provided in a conference abstract: there was no deterioration in role function, global QL, body image, or sexual function | 30 months | |
| FACT‐General and FACT‐B | Health‐related QoL mean scores were worse in the OFS group than in the no OFS group, based on scores from the FACT‐General and FACT‐B cancer subscales at all time points (i.e. 6 months, 12 months, years 2, 3, 4, and 5). This was more pronounced over time and reached statistical and clinical significance for FACT‐G at 3 years. This statistical difference did not persist at 4 and 5 years, and decreased over time | 5.86 years | |
| International Breast Cancer Study Group (IBCSG) QoL core form and a symptom‐specific module | 1722 out of 2045 participants were tested at baseline, at 6, 12, 18, and 24 months, and then annually during 3 to 6 years. Baseline global QoL scores were similar in both arms and were similar between treatments over the whole treatment period The OFS group reported worse endocrine symptoms and sexual functioning during the first 2 years of treatment. In particular, a decline in sexual interest was significantly greater among participants in the OFS group at 6 months compared to the no OFS group, but not at 24 and 60 months. Participants in the OFS group also experienced worse vaginal dryness over the whole treatment period, while participants in the no OFS group experienced a significantly greater increase in vaginal discharge and itching in the short and intermediate term. Both groups had worsening of bone and joint pain and weight gain. Participants in the OFS arm had more sleep disturbance at 6 months than those receiving tamoxifen alone but not at later time points. No significant differences between groups were observed for all other symptoms. For example, changes in other symptoms (headache, irritable, feeling dizzy, appetite, feeling sick, tired) and global indicators for physical well‐being, mood, and health perception were small over time and were similar between groups The OFS group had slightly less improvement in coping effort and were more burdened by treatment at 6 and 24 months than those in the no OFS group | 6 years | |
| NR | NR | NR | |
| NR | NR | NR | |
| Assessed using 7 domains: daily activity, appetite, sleep, mental, economic, menopause symptoms and menstruation. No further details provided | None of the scores changed considerably before administration of and after treatments | 2 years | |
| OFS + chemotherapy vs chemotherapy | |||
| NR | NR | NR | |
| NR | NR | NR | |
| NR | NR | NR | |
| NR | NR | NR | |
| QoL IBCSG core questionnaire | QoL scores were similar across ER‐positive and ER‐negative cohorts (however data were not shown). There were no differences between groups except for hot flushes at 3 years. Quality of life measures were taken at baseline, 3 months, 6 months, and 36 months. The Discussion section of the trial publication stated that "CMF followed by goserelin showed the same effect on all QoL indicators as CMF alone" p.269 | 3 years after randomisation | |
| NR | NR | NR | |
| OFS + chemotherapy + tamoxifen vs chemotherapy + tamoxifen | |||
| NR | NR | NR | |
| NR | NR | NR | |
| NR: not reported (meaning not measured). | |||
| Study | Questionnaire used | Summary of findings | Follow‐up period |
| OFS versus observation | |||
| NR | NR | NR | |
| OFS + tamoxifen vs tamoxifen | |||
| EORTC QLQ‐C30, BR23 breast cancer module | 246 out of 2144 participants agreed to take part in the Quality of Life substudy. A narrative synthesis of the key findings was provided in a conference abstract: there was no deterioration in role function, global QL, body image, or sexual function | 30 months | |
| FACT‐General and FACT‐B | Health‐related QoL mean scores were worse in the OFS group than in the no OFS group, based on scores from the FACT‐General and FACT‐B cancer subscales at all time points (i.e. 6 months, 12 months, years 2, 3, 4, and 5). This was more pronounced over time and reached statistical and clinical significance for FACT‐G at 3 years. This statistical difference did not persist at 4 and 5 years, and decreased over time | 5.86 years | |
| International Breast Cancer Study Group (IBCSG) QoL core form and a symptom‐specific module | 1722 out of 2045 participants were tested at baseline, at 6, 12, 18, and 24 months, and then annually during 3 to 6 years. Baseline global QoL scores were similar in both arms and were similar between treatments over the whole treatment period The OFS group reported worse endocrine symptoms and sexual functioning during the first 2 years of treatment. In particular, a decline in sexual interest was significantly greater among participants in the OFS group at 6 months compared to the no OFS group, but not at 24 and 60 months. Participants in the OFS group also experienced worse vaginal dryness over the whole treatment period, while participants in the no OFS group experienced a significantly greater increase in vaginal discharge and itching in the short and intermediate term. Both groups had worsening of bone and joint pain and weight gain. Participants in the OFS arm had more sleep disturbance at 6 months than those receiving tamoxifen alone but not at later time points. No significant differences between groups were observed for all other symptoms. For example, changes in other symptoms (headache, irritable, feeling dizzy, appetite, feeling sick, tired) and global indicators for physical well‐being, mood, and health perception were small over time and were similar between groups The OFS group had slightly less improvement in coping effort and were more burdened by treatment at 6 and 24 months than those in the no OFS group | 6 years | |
| NR | NR | NR | |
| NR | NR | NR | |
| Assessed using 7 domains: daily activity, appetite, sleep, mental, economic, menopause symptoms and menstruation. No further details provided | None of the scores changed considerably before administration of and after treatments | 2 years | |
| OFS + chemotherapy vs chemotherapy | |||
| NR | NR | NR | |
| NR | NR | NR | |
| NR | NR | NR | |
| NR | NR | NR | |
| QoL IBCSG core questionnaire | QoL scores were similar across ER‐positive and ER‐negative cohorts (however data were not shown). There were no differences between groups except for hot flushes at 3 years. Quality of life measures were taken at baseline, 3 months, 6 months, and 36 months. The Discussion section of the trial publication stated that "CMF followed by goserelin showed the same effect on all QoL indicators as CMF alone" p.269 | 3 years after randomisation | |
| NR | NR | NR | |
| OFS + chemotherapy + tamoxifen vs chemotherapy + tamoxifen | |||
| NR | NR | NR | |
| NR | NR | NR | |
| NR: not reported (meaning not measured). | |||
