Vasodilatadores para el fenómeno de Raynaud primario
Resumen
Antecedentes
Se han sugerido muchos fármacos para el tratamiento sintomático del fenómeno de Raynaud primario. Aparte de los bloqueadores de los canales de calcio, que se consideran los fármacos de elección, la evidencia de los efectos de los tratamientos farmacológicos alternativos es limitada. Ésta es una actualización de una revisión publicada por primera vez en 2008.
Objetivos
Evaluar los efectos de los fármacos con efectos vasodilatadores en el fenómeno de Raynaud primario según la frecuencia, la intensidad y la duración de los episodios vasoespásticos; la calidad de vida; los eventos adversos y la puntuación de la Raynauds Condition Score.
Métodos de búsqueda
El documentalista del Grupo Cochrane Vascular (Cochrane Vascular) buscó en el Registro especializado del Grupo Cochrane Vascular, en las bases de datos CENTRAL, MEDLINE, Embase y CINAHL y en los registros de ensayos de la Plataforma de registros internacionales de ensayos clínicos de la Organización Mundial de la Salud y ClinicalTrials.gov hasta el 16 de noviembre de 2020.
Criterios de selección
Se incluyeron los ensayos controlados aleatorizados que evaluaran los efectos de las formulaciones orales, intravenosas y tópicas de cualquier fármaco con efectos vasodilatadores sobre los síntomas subjetivos, las puntuaciones de intensidad y los desenlaces radiológicos del fenómeno de Raynaud primario. En esta revisión no se evaluó el tratamiento con bloqueadores de los canales de calcio ni se compararon estos fármacos.
Obtención y análisis de los datos
Dos autores de la revisión seleccionaron de forma independiente los estudios para inclusión, los evaluaron con la herramienta "Risk of bias" de Cochrane y extrajeron los datos de los estudios. Los desenlaces de interés incluyeron la frecuencia, la intensidad y la duración de los episodios; la calidad de vida (CdV); los eventos adversos (EA); y la puntuación de la Raynaud Condition Score (RCS). La certeza de la evidencia se evaluó mediante el método GRADE.
Resultados principales
En esta actualización se identificaron siete estudios nuevos. En total se incluyeron 15 estudios con 635 participantes. Estos estudios compararon diferentes vasodilatadores con placebo. Los estudios individuales utilizaron distintos métodos y medidas para informar sobre los diferentes desenlaces.
Inhibidores de la enzima convertidora de angiotensina (IECA)
La combinación de los datos de tres estudios reveló un posible pequeño aumento de la frecuencia de episodios por semana después del tratamiento (captopril o enalapril) en comparación con el placebo (diferencia de medias [DM] 0,79; intervalo de confianza [IC] del 95%: 0,43 a 1,17; evidencia de certeza baja). No hubo evidencia de una diferencia entre los grupos en cuanto a la intensidad de los episodios (DM ‐0,17; IC del 95%: ‐4,66 a 4,31; 34 participantes, dos estudios; evidencia de certeza baja), la duración de los episodios (DM 0,54; IC del 95%: ‐2,42 a 1,34; 14 participantes, un estudio; evidencia de certeza baja) o los EA (razón de riesgos [RR] 1,35; IC del 95%: 0,67 a 2,73; 46 participantes, tres estudios; evidencia de certeza baja). No se informó la CdV ni la puntuación RCS.
Alfabloqueantes
Dos estudios utilizaron alfabloqueantes (buflomedil o moxisilita). No fue posible combinar los datos debido a la manera en que se presentaron los resultados. Es probable que el buflomedil redujera la frecuencia de los episodios comparado con el placebo (DM ‐8,82; IC del 95%: ‐11,04 a ‐6,60; 31 participantes, un estudio; evidencia de certeza moderada) y podría mejorar las puntuaciones de intensidad (DM ‐0,41; IC del 95%: ‐0,62 a ‐0,30; evidencia de certeza moderada). Con la moxisilita, los investigadores notificaron un menor número de episodios (p < 0,02), síntomas menos graves (p < 0,01) y una menor duración de los episodios, pero la relevancia clínica de estos resultados no está clara. No se detectó evidencia de diferencias en los EA entre los grupos de buflomedil y placebo (RR 1,41; IC del 95%: 0,27 a 7,28; 31 participantes, un estudio; evidencia de certeza moderada). Se observaron más EA en los participantes del grupo de moxisilita que en los del grupo placebo.
Análogos de la prostaciclina/prostaglandina
Un estudio comparó beraprost versus placebo. No hubo evidencia de un beneficio en la frecuencia (DM 2,00; IC del 95%: ‐0,35 a 4,35; 118 participantes; evidencia de certeza baja) ni en la intensidad (DM ‐0,06; IC del 95%: ‐0,34 a 0,22; 118 participantes; evidencia de certeza baja) de los episodios. En general, se detectaron más EA en el grupo de beraprost (RR 1,59; IC del 95%: 1,05 a 2,42; 125 participantes; evidencia de certeza baja). Este estudio no informó sobre la duración de los episodios, la CdV ni la puntuación RCS.
Inhibidores de la tromboxano sintetasa
Un estudio comparó un inhibidor de la tromboxano sintetasa (dazoxiben) versus placebo. No hubo evidencia de un beneficio en la frecuencia de los episodios (DM 0,8; IC del 95%: ‐1,81 a 3,41; seis participantes; evidencia de certeza muy baja). No se informó sobre los eventos adversos en los análisis de subgrupos de los participantes con fenómeno de Raynaud primario, y el estudio no informó la duración de los episodios, la intensidad de los síntomas, la CdV ni la puntuación RCS.
Inhibidores selectivos de la recaptación de serotonina
Un estudio comparó la ketanserina con placebo. Podría haber una ligera reducción del número de episodios por semana con la ketanserina comparada con el placebo (DM ‐14,0; IC del 95%: ‐27,72 a ‐0,28; 41 participantes; evidencia de certeza muy baja) y una reducción en las puntuaciones de intensidad (DM ‐133,00; IC del 95%: ‐162,40 a ‐103,60; 41 participantes; evidencia de certeza muy baja). No hubo evidencia de que la ketanserina redujera la duración de los episodios (DM ‐4,00; IC del 95%: ‐14,82 a 6,82; 41 participantes; evidencia de certeza muy baja) ni de que los EA aumentaran en ninguno de los grupos (RR 1,54; IC del 95%: 0,89 a 2,65; 41 participantes; evidencia de certeza muy baja). Este estudio no informó sobre la CdV ni la puntuación RCS.
Nitrato/derivados de nitrato
Cuatro estudios compararon tratamientos tópicos de nitroglicerina o trinitrato de glicerilo versus placebo, cada uno de los cuales informó sobre pocos desenlaces. El metanálisis mostró que no hubo evidencia de un efecto sobre la frecuencia de los episodios por semana (DM ‐1,57; IC del 95%: ‐4,31 a 1,17; 86 participantes, dos estudios; evidencia de certeza muy baja). No fue posible agrupar ninguno de los datos de los otros desenlaces.
Inhibidores de la fosfodiesterasa
Tres estudios compararon los inhibidores de la fosfodiesterasa (vardenafil, cilostazol o PF‐00489791) con un placebo equivalente. Los resultados mostraron que no hubo evidencia de una diferencia en cuanto a la frecuencia de los episodios (DM estandarizada [DME] ‐0,05; IC del 95%: ‐6,71 a 6,61; 111 participantes, dos estudios; evidencia de certeza baja), la intensidad de los episodios (DM ‐0,03; IC del 95%: ‐1,04 a 0,97; 111 participantes, dos estudios; evidencia de certeza muy baja), la duración de los episodios (DM ‐1,60; IC del 95%: ‐7,51 a 4,31; 73 participantes, un estudio; evidencia de certeza baja) o la puntuación RCS (DME ‐0,8; IC del 95%: ‐1,74 a 0,13; 79 participantes, dos estudios; evidencia de certeza baja). Los autores de los estudios informaron que el 35% de participantes que recibieron cilostazol notificaron cefaleas, las cuales no se notificaron en el grupo placebo. El PF‐00489791 causó que 34 de 54 participantes experimentaran EA versus 43 de 102 participantes que recibieron placebo (RR 1,49). La cefalea fue el más frecuente, que afectó a 14 participantes (PF‐00489791) versus nueve participantes (placebo).
Conclusiones de los autores
Los estudios incluidos investigaron varios vasodilatadores diferentes (tópicos y orales) para el tratamiento del fenómeno de Raynaud primario. El pequeño tamaño muestral, la limitación de los datos y la variabilidad en el informe de los desenlaces dieron lugar a evidencia de certeza muy baja a moderada. La evidencia no es suficiente para apoyar el uso de los vasodilatadores e indica que su uso podría incluso empeorar la enfermedad.
PICO
Resumen en términos sencillos
Medicamentos vasodilatadores para reducir los síntomas del fenómeno de Raynaud primario
Antecedentes
El fenómeno de Raynaud (FR) es una enfermedad que afecta a los pequeños vasos sanguíneos de las extremidades, normalmente en los dedos de las manos, pero también en los dedos de los pies y otras partes del cuerpo. Está causada por un estrechamiento temporal de los vasos sanguíneos, que provoca cambios de color con entumecimiento, hormigueo y dolor asociados. Se conocen varios factores desencadenantes de esta afección, como el estrés, el frío y el uso de herramientas manuales vibratorias. Las medidas conservadoras para controlar esta enfermedad incluyen dejar de fumar y mantener el calor periférico y ambiental. Pueden utilizarse medicamentos que dilatan los vasos sanguíneos, como los bloqueadores de los canales de calcio (BCC), pero pueden tener efectos secundarios. Esta revisión trata de examinar la efectividad y la seguridad de medicamentos dilatadores de los vasos sanguíneos que no fueran BCC.
Características de los estudios y resultados clave
En esta actualización se encontraron siete estudios nuevos, por lo que el total de estudios es 15 (la búsqueda se actualizó el 16 de noviembre de 2020). Esta actualización incluye ahora otras vías de administración como por vena (intravenosa) y por la piel (tópica), a diferencia de las revisiones anteriores, que se centraban en las formas orales de tratamiento. Los estudios se publicaron entre 1989 y 2013 e implicaron a 635 participantes asignados al azar a recibir tratamiento o control con placebo. Muchos ensayos no describieron varios aspectos de la metodología de los estudios como la asignación al azar, la ocultación de la asignación y el cegamiento. La duración del tratamiento varió entre dos semanas y seis meses.
Los inhibidores de la enzima convertidora de angiotensina (IECA) como clase de medicamento, en concreto el enalapril y el captopril en esta revisión, en general aumentaron la frecuencia de los episodios de Raynaud a la semana, pero no afectaron la intensidad de los mismos. El enalapril empeoró la evaluación subjetiva de la mejoría, y el captopril no mejoró los desenlaces subjetivos ni el flujo sanguíneo digital. El buflomedil mostró una ligera reducción de la frecuencia y la intensidad de los episodios con un aumento de los efectos secundarios. El beraprost y el dazoxiben no mostraron cambios en la frecuencia o la intensidad de los episodios ni en la puntuación de la discapacidad y se asociaron a un aumento de los efectos secundarios. La ketanserina no mostró una mejoría de la frecuencia o la duración de los episodios ni del flujo sanguíneo digital, pero sí en la puntuación de la intensidad del FR. En un estudio pequeño, la moxisilita mostró reducir ligeramente la frecuencia y la intensidad de los episodios, pero con un aumento de los efectos secundarios. El trinitrato de glicerilo tópico no mostró efectos de reducción de la frecuencia de los episodios por semana. Un estudio informó una mejoría subjetiva en la puntuación de la escala Raynaud Condition Score (RCS). Un pequeño estudio informó mejorías en la frecuencia y la intensidad de los episodios como cambios subjetivos. El dolor de cabeza fue el efecto secundario más significativo y frecuente del tratamiento. Los inhibidores de la fosfodiesterasa no provocaron una reducción de la frecuencia, la intensidad ni la duración de los episodios ni mejoraron la puntuación de la RCS. Un estudio informó a favor del vardenafil por si solo para reducir la puntuación RCS, pero encontró que el efecto es probablemente pequeño. Un estudio informó que el cilostazol aumentó la frecuencia y la intensidad de los episodios; se necesitan más estudios para confirmar este hallazgo. El riesgo de dolor de cabeza como efecto secundario del tratamiento aumentó con el uso de cilostazol. Se observó que el PF‐00489791 a una dosis de 20 mg mejoró ligeramente todos los criterios de valoración subjetivos y la puntuación RCS.
Fiabilidad de la evidencia
La confianza en estos resultados es de muy baja a moderada, por lo que no es posible establecer conclusiones firmes sobre el beneficio de estos medicamentos en la mejoría de los síntomas del FR primario. No existe seguridad debido a la escasa cantidad de participantes en los estudios, los problemas con la forma en que los estudios fueron diseñados y las diferencias en la manera de medir si los tratamientos fueron o no eficaces. Por lo tanto, la importancia clínica de los resultados es difícil de evaluar, especialmente si la respuesta al placebo es alta. También es pertinente que los resultados para cada clase de medicamento, y para los medicamentos individuales dentro de cada clase, deben interpretarse en el contexto de que podrían tener efectos farmacológicos variables además de la vasodilatación. Esto debe tenerse en cuenta a la hora de establecer cualquier conclusión sobre los efectos globales de cada clase de medicamento o cada medicamento individual.
Authors' conclusions
Summary of findings
| ACE inhibitors compared to placebo for treatment of primary Raynaud's phenomenon | ||||||
| Patient or population: primary Raynaud's phenomenon | ||||||
| Outcomes | Anticipated absolute effects* (95% CI) | Relative effect | №. of participants | Certainty of the evidence | Comments | |
|---|---|---|---|---|---|---|
| Risk with placebo | Risk with ACE inhibitor | |||||
| Frequency of attacks (per week) (4 to 6‐week follow‐up) | MD 0.79 higher | ‐ | 44 (3 RCTs) | ⊕⊕⊝⊝ | There may be increased frequency of attacks with treatment | |
| Severity of attacks (mild, moderate, and severe. then converted to numeric representation) (4 to 6‐week follow‐up) | MD 0.17 lower | ‐ | 34 (2 RCTs) | ⊕⊕⊝⊝ | No evidence of a difference | |
| Duration of attacks (minutes) (4 to 6‐week follow‐up) | MD 0.54 higher (1.34 lower to 2.42 higher) | ‐ | 14 (1 RCT) | ⊕⊕⊝⊝ | No evidence of a difference | |
| QoL | See comment | ‐ | ‐ | ‐ | This outcome was not reported by any study | |
| Adverse events (4 to 6‐week follow‐up) | 182 per 1000 | 245 per 1000 (122 to 496) | RR 1.35 (95% CI 0.67 to 2.73) | 46 (3 RCTs) | ⊕⊕⊝⊝ | No evidence of a difference |
| RCS | See comment | ‐ | ‐ | ‐ | This outcome was not reported by any study | |
| *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. | ||||||
| aOutpatient clinic ‐ Madsen 1984 and Challenor 1991 did not specify. | ||||||
| Alpha blockers compared to placebo for primary Raynaud's phenomenon | ||||||
| Patient or population: primary Raynaud's phenomenon | ||||||
| Outcomes | Anticipated absolute effects* (95% CI) | Relative effect | №. of participants | Certainty of the evidence | Comments | |
|---|---|---|---|---|---|---|
| Risk with placebo | Risk with alpha blocker | |||||
| Frequency of attacks (4 weeks' to 180 days' follow‐up) | Moxisylyte | ‐ | 33 (1 RCT) | ‐ | 19 participants had fewer attacks during the moxisylyte period and 10 during the placebo period. Four participants had an equal number of attacks in each period | |
| See comment | ||||||
| Buflomedil | ‐ | 31 (1 RCT) | ⊕⊕⊕⊝ | Frequency of attacks per week may be reduced with buflomedil treatment | ||
| MD 8.82 lower | ||||||
| Severity of attacks (Dichotomous outcome over 4 weeks' to 180 days' follow‐up) | Moxisylyte | ‐ | 25 (1 RCT) | ‐ | Of 25 participants, 7 reported more severe attacks during the moxisylyte period and 18 participants reported more severe attacks during the placebo period | |
| See comment | ||||||
| Buflomedil | ‐ | 31 (1 RCT) | ⊕⊕⊝⊝ | Severity of attacks may be slightly reduced | ||
| MD 0.41 lower | ||||||
| Duration of attacks (4 weeks) | Moxisylyte | ‐ | 33 | ‐ | Fifteen participants recorded shorter total duration of attacks while on moxisylyte, and 9 had shorter duration of attacks on placebo | |
| See comment | ||||||
| Buflomedil | ‐ | ‐ | ‐ | This outcome was not reported | ||
| See comment | ||||||
| QoL | See comment | ‐ | ‐ | ‐ | This outcome was not reported | |
| Adverse events (4 weeks' to 180 days' follow‐up) | Moxisylyte | ‐ | 33 | ‐ | Total of 13 participants during the moxisylyte phase and 3 during the placebo phase reported adverse events | |
| See comment | ||||||
| Buflomedil | RR 1.41 | 31 (1 RCT) | ⊕⊕⊕⊝ | No evidence of a difference in adverse events was seen between buflomedil and placebo groups | ||
| 133 per 1000 | 188 per 1000 | |||||
| RCS | See comment | ‐ | ‐ | ‐ | This outcome was not reported | |
| *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. | ||||||
| aBuflomedil (in Le Quentrec 1991) and moxisylyte (thymoxamine; in Jaffe 1980). | ||||||
| Beraprost compared to placebo for primary Raynaud's phenomenon | ||||||
| Patient or population: primary Raynaud's phenomenon | ||||||
| Outcomes | Anticipated absolute effects* (95% CI) | Relative effect | №. of participants | Certainty of the evidence | Comments | |
|---|---|---|---|---|---|---|
| Risk with placebo | Risk with beraprost | |||||
| Frequency of attacks per week (6 weeks' follow‐ up) | MD 2 higher | ‐ | 118 (1 RCT) | ⊕⊕⊝⊝ | No evidence of a difference | |
| Severity of attacks (1 to 4 scale, 6 weeks' follow‐up) | MD 0.06 lower | ‐ | 118 | ⊕⊕⊝⊝ | No evidence of a difference | |
| Duration of attacks | See comment | ‐ | ‐ | ‐ | This outcome was not reported | |
| QoL | See comment | ‐ | ‐ | ‐ | This outcome was not reported | |
| Adverse events (6 weeks' follow‐up) | Study population | RR 1.59 | 125 | ⊕⊕⊝⊝ | There may be more adverse events in the beraprost group | |
| 339 per 1000 | 539 per 1000 | |||||
| RCS | See comment | This outcome was not reported | ||||
| *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. | ||||||
| aWe downgraded by one step due to inconsistency. | ||||||
| Dazoxiben compared to placebo for primary Raynaud's phenomenon | ||||||
| Patient or population: primary Raynaud's phenomenon | ||||||
| Outcomes | Anticipated absolute effects* (95% CI) | Relative effect | №. of participants | Certainty of the evidence | Comments | |
|---|---|---|---|---|---|---|
| Risk with placebo | Risk with dazoxiben | |||||
| Frequency of attacks Per week (over 2 weeks) | MD 0.8 higher | ‐ | 6 (1 RCT) | ⊕⊝⊝⊝ | There was no evidence of an effect of dazoxiben compared with placebo | |
| Severity of attacks | See comment | ‐ | ‐ | ‐ | This outcome was not reported | |
| Duration of attacks | See comment | ‐ | ‐ | ‐ | This outcome was not reported | |
| QoL | See comment | ‐ | ‐ | ‐ | This outcome was not reported | |
| Adverse events | See comment | ‐ | ‐ | ‐ | Unclear how many participants had primary Raynaud's phenomenon among 5 participants who had adverse events | |
| RCS | See comment | ‐ | ‐ | ‐ | This outcome was not reported | |
| *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. | ||||||
| aWe downgraded by three steps due to risk of bias concerns (selection and performance bias), inconsistency of result, and imprecision (small number of participants with primary Raynaud's phenomenon). | ||||||
| Ketanserin compared to placebo for primary Raynaud's phenomenon | ||||||
| Patient or population: primary Raynaud's phenomenon | ||||||
| Outcomes | Anticipated absolute effects* (95% CI) | Relative effect | №. of participants | Certainty of the evidence | Comments | |
|---|---|---|---|---|---|---|
| Risk with placebo | Risk with ketanserin | |||||
| Frequency of attacks Per week (over 4 weeks) | MD 14 lower | ‐ | 41 (1 RCT) | ⊕⊝⊝⊝ | There may be a slight reduction in the number of attacks per week in the ketanserin group compared to the placebo group | |
| Severity of attacks (frequency of attacks/d × duration of attacks, over 4 weeks) | MD 133 lower | ‐ | 41 (1 RCT) | ⊕⊝⊝⊝ | Severity score may be slightly reduced after ketanserin compared to placebo | |
| Duration of attacks Per day (minutes, over 4 weeks) | MD 4 lower | ‐ | 41 (1 RCT) | ⊕⊝⊝⊝ | No evidence of a difference in duration of attacks | |
| QoL | See comment | ‐ | ‐ | ‐ | This outcome was not reported | |
| Adverse events (over 4 weeks) | Study population | RR 1.54 | 41 | ⊕⊝⊝⊝ | Headache, dry mouth. and dizziness were reported more frequently in the treatment group | |
| 317 per 1000 | 488 per 1000 | |||||
| RCS | See comment | ‐ | ‐ | ‐ | This outcome was not reported | |
| *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. | ||||||
| aWe downgraded by three steps due to risk of bias concerns (selection and performance bias), imprecision (small number of participants), and inconsistency (wide confidence intervals). | ||||||
| Nitroglycerin or glyceryl trinitrate (GTN) compared to placebo for primary Raynaud's phenomenon | ||||||
| Patient or population: primary Raynaud's phenomenon | ||||||
| Outcomes | Anticipated absolute effects* (95% CI) | Relative effect | №. of participants | Certainty of the evidence | Comments | |
|---|---|---|---|---|---|---|
| Risk with placebo | Risk with GTN | |||||
| Frequency of attacks (1 or 4 weeks' follow‐up) | Continuous data | ‐ | 86 (2 RCTs) | ⊕⊝⊝⊝ | No evidence of a difference | |
| MD 1.57 lower | ||||||
| Dichotomous data | ‐ | 14 | ‐ | Six participants reported response to Nitroderm compared to 1 in the placebo group | ||
| See comment | ||||||
| Severity of attacks (1 week follow‐up) | Continuous data | ‐ | 17 (1 RCT) | ⊕⊝⊝⊝ | Severity may be reduced but the clinical relevance of this is unclear | |
| MD 4.25 lower | ||||||
| Dichotomous data | ‐ | 14 | ‐ | Five participants reported a positive result compared to 1 in the placebo group | ||
| See comment | ||||||
| Duration of attacks (2 or 4 weeks' follow‐up) | See comment | ‐ | 77 (2 RCTs) | ‐ | Sovijarvi 1984 reported lack of differences in duration of attacks between GTN and placebo groups (no data provided) Chung 2009 reported no significant decrease in duration of attacks | |
| QoL | See comment | ‐ | ‐ | ‐ | No studies reported on this outcome | |
| Adverse events (3 weeks' follow‐up) | See comment | ‐ | 8 | ‐ | All 8 participants reported headaches with nitroglycerin | |
| RCS (4 weeks' follow‐up) | MD 0.36 lower | ‐ | 69 (1 RCT) | ⊕⊕⊝⊝ | No evidence of a difference following GTN treatment | |
| *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. | ||||||
| aStudies were conducted in outpatient settings, except for Sovijarvi 1984 (laboratory room). | ||||||
| Phosphodiesterase inhibitors compared to placebo for primary Raynaud's phenomenon | ||||||
| Patient or population: primary Raynaud's phenomenon | ||||||
| Outcomes | Anticipated absolute effects* (95% CI) | Relative effect | №. of participants | Certainty of the evidence | Comments | |
|---|---|---|---|---|---|---|
| Risk with placebo | Risk with phosphodiesterase inhibitors | |||||
| Frequency of attacks (per week) (4 and 6 weeks' follow‐up) | SMD 0.05 lower | ‐ | 111 (2 RCTs) | ⊕⊕⊝⊝ | No evidence of a difference but studies showed conflicting results | |
| Severity of attacks (Likert scale 1 to 9 or 11 points; 4 and 6 weeks' follow‐up) | SMD 0.03 lower | ‐ | 111 (2 RCTs) | ⊕⊝⊝⊝ | No evidence of a difference but studies showed conflicting results | |
| Duration of attacks (4 weeks' follow‐up) | MD 1.60 lower (7.51 lower to 4.31 higher) | ‐ | 73 (1 RCT) | ⊕⊕⊝⊝ | No evidence of a difference | |
| QoL | See comment | ‐ | ‐ | ‐ | This outcome was not reported | |
| Adverse events | See comment | ‐ | 138 (3 RCTs) | ‐ | 35% of participants on cilostazol complained of headaches. Two participants on cilostazol complained of palpitations These were not reported in the placebo group No specific adverse events were reported in the vardenafil group compared to the placebo group 34/54 participants experienced adverse events in the PF‐00489791 treatment group compared with 43/102 participants in the placebo group. Headache was the most commonly reported adverse event, affecting 14 participants in the PF‐00489791 group and 9 participants in the placebo group | |
| RCS (6 weeks' follow‐up) | SMD 0.80 lower | ‐ | 79 (2 RCTs) | ⊕⊕⊝⊝ | No evidence of a difference | |
| *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. | ||||||
| aStudies investigated cilostazol (Rajagopalan 2003); vardenafil (Caglayan 2012); and PF‐00489791 (NCT01090492). | ||||||
Background
Description of the condition
The phenomenon of episodic digital ischemia provoked by cold, cyanosis, and emotion was first described by Maurice Raynaud in 1862 (Belch 1990). At average onset of 40 years of age (Pope 2011), this common condition is estimated to have a prevalence of 1.6% to 7.2% in the general population, more likely affecting women at 2.1% to 15.8%, compared to 0.8% to 6.5% among men (Garner 2015).
Raynaud's phenomenon (RP) is classically characterized by "triphasic" color change, reflecting the vasospastic nature of the condition. Observed changes often take place in three distinct phases (Hughes 2016). The first phase, reflective of vasospasm/constriction of the small arteries of the distal digits, leads to the well‐circumscribed pallor of the fingers. The second phase is caused by cyanosis, which quickly resolves and followed by the final phase involving hyperemia, characterized by restoration of blood flow (Bakst 2008).
Raynaud's phenomenon is divided into primary, also known as idiopathic disease, and secondary, which is associated with various disease states. It usually occurs in the presence of triggers such as stress or cold and in occupations that utilize heavy or vibrational machinery. Raynaud's phenomenon commonly affects the fingers and toes but has been observed to affect other areas such as ears, nose, and even nipples (Maverakis 2014; Prodigy 2006). Common rheumatologic conditions that can predispose include systemic sclerosis, systemic lupus erythematosus, and vasculitides. Certain medications, including beta blockers, ergot alkaloids, and chemotherapeutics such as cisplatin and bleomycin, can also act as triggers (Wigley 2016).
The exact pathophysiology of RP is not completely understood but is postulated to involve a complex interplay of factors that impair vasoconstriction and vasodilatation, as well as intravascular factors. Vasoconstriction is thought to be mediated by alpha‐2 adrenergic effects and to be more sensitive in the distal vascular beds and more responsive to a cold stimulus (Cooke 2005). Vasoconstriction is also increased among patients with systemic sclerosis. Vasodilatory effects are thought to be related to calcitonin gene‐related peptide, but many other factors such as substance P, neurokinin A, and vasoactive intestinal peptide are thought to be implicated. It has been observed that RP can be exacerbated by conditions that affect the viscosity of blood vessels; suggested intravascular factors include platelet activation and oxidative stress. Finally, genetic and hormonal factors, in addition to neural abnormalities of vessels, have been implicated in the pathogenesis (Prete 2014).
The approach to diagnosis of RP has been updated since publication of the previous review, and agreement on these changes was reached by an international consensus panel in 2013 (Maverakis 2014). The new diagnostic criteria require a patient to be diagnosed with RP in three steps, yielding a disease score. The condition is then classified into primary disease based on a normal capillaroscopy, a physical examination negative for findings to suggest a secondary cause, the absence of a history of connective tissue disease, and a negative or low‐titer antinuclear antibody (such as 1:40) by indirect immunofluorescence. Notably, a negative erythrocyte sedimentation rate is no longer required (Maverakis 2014).
Description of the intervention
Symptomatic treatment for primary RP consists of conservative measures such as avoiding exposure to cold and using protective clothing. Smoking and other behaviors that may contribute to symptoms including use of drugs or vibratory tools should be avoided. In comparison to those with secondary RP, in which trophic changes, painful ulcers, and gangrene may develop, requiring more invasive measures, most patients with primary RP can control their symptoms by using conservative measures (Block 2001). If attacks persist despite adequate conservative measures, use of a calcium channel blocker (CCB) is recommended as first‐line treatment (Pope 2011; Prodigy 2006). Treatment with CCBs is not discussed in this review, as it is addressed in another Cochrane systematic review (Ennis 2016). Several side effects have been reported with use of CCBs, and a range of other drugs have been used in multiple studies owing to their vasodilatory action. A list of drugs with vasodilatory effects is included in Table 1 (AMH 2018). Medications used primarily for conditions such as hypertension, depression, ischemic heart disease, secondary RP, and erectile dysfunction are also used in treating primary RP, with only limited data available regarding their effectiveness. These include the alpha blockers buflomedil and moxisylyte (thymoxamine), the thromboxane synthase inhibitor dazoxiben, and PF‐00489791, a phosphodiesterase Inhibitor that has been used only in early trial stages.
| Pharmacologic class | ATC code | Generic name |
|---|---|---|
| Phosphodiesterase‐3 inhibitor | C01CE02, B01AC23 | Milrinone, cilostazol |
| Neprilysin inhibitor | C09DX04 | Sacubitril |
| Nitrate/nitrate derivative | C01D, C08EX02, C02DD01 | Glyceryl trinitrate, isosorbide dinitrate, isosorbide mononitrate, nicorandil, sodium nitroprusside |
| Angiotensin‐converting enzyme (ACE) inhibitor | C09A | Captopril, enalapril, fosinopril, lisinopril, perindopril, quinapril, ramipril, trandolapril |
| Angiotensin‐II receptor antagonist | C09C | Candesartan, eprosartan, irbesartan, losartan, olmesartan, telmisartan, valsartan |
| Selective alpha‐blocker | C02CA, C02AC | Prazosin, terazosin, moxonidine, tamsulosin, alfuzosin |
| Potassium channel activator | C02DA | Diazoxide |
| Reflex sympathetic stimulator | C02D | Hydralazine, minoxidil |
| Prostacyclin/prostanoid | B01AC, G04BE01 | Epoprostenol, iloprost, treprostinil, alprostadil, beraprost |
| Endothelin antagonist | C02KX | Ambrisentan, bosentan, macitentan, rociguat |
| Phosphodiesterase‐5 inhibitor | G04BE | Sildenafil, tadalafil, vardenafil |
| Selective serotonin reuptake inhibitor | N06AB, C02KD01 | Citalopram, dapoxetine, escitalopram, fluoxetine, fluvoxamine, paroxetine, sertraline, ketanserin |
| Peripheral vasodilator, mechanism not well defined | C04AD | Pentoxifylline, perhexiline |
AMH: Australian Medicines Handbook.
ATC: Anatomical Therapeutic Chemical Classification.
How the intervention might work
The pathophysiology of RP involves intermittent vasodilation and vasoconstriction, resulting in episodic digital ischemia. It is therefore the action of vasodilatory agents such as CCBs that provides at least moderate effectiveness in this condition (Smith 1985). Other interventions with vasodilatory properties have not yet been fully validated and include oral, topical, and intravenous medications that may be administered locally to affected digits and/or systemically. These include angiotensin‐converting enzyme (ACE) inhibitors, nitrates, alpha blockers, phosphodiesterase inhibitors, prostacyclins, and selective serotonin receptor inhibitors (SSRIs). These interventions have worked with other vasospastic conditions including coronary artery vasospasm and cerebral artery vasospasm and in related conditions such as erectile dysfunction and pulmonary arterial hypertension. For example, ACE inhibitors and angiotensin receptor blockers (ARBs) are widely used for high blood pressure, nitrates are used in acute myocardial infarction to dilate coronary vessels, and prostacyclins and phosphodiesterase‐5 inhibitors may be used in primary pulmonary hypertension (Duarte 2013).
Antithrombotic treatments have been used in severe disease, especially in secondary RP as the result of frequent complications such as severe ischemia leading to digital amputation, but less often in primary RP. Regulation of vascular tone involves many factors and defects in the complex interaction between smooth muscle and endothelium, and innervation of vessels contributes to primary RP. This leads to abnormal autoregulation of the small blood vessels causing vasoconstriction and is thought to be the primary defect in primary RP. The underlying molecular process is poorly understood but has been postulated to be caused by a combination of vasoconstrictive mediators, exaggerated vasoconstrictive responses, and blunted vasodilatory responses (Block 2001). Angiotensin‐II, a vasoconstrictor once bound to angiotensin‐II receptors on vascular smooth muscle, can be inhibited by ACE inhibitors or by angiotensin‐II receptor antagonists. The sympathetic nervous system releases catecholamine, which binds to alpha receptors, resulting in vasoconstriction of peripheral vessels that would be blocked by alpha blockers. Alpha blockers produce their effects by blocking alpha receptors that constrict vessels mediated by catecholamines. Serotonin released by platelets contributes to vasospasm in scleroderma‐related RP, leading to the postulated effects of serotonergic S2 receptor antagonists and serotonin‐converting enzyme inhibitors as vasodilators. Nitrate and its derivatives produce vasodilatory effects through donation of nitric oxide, which results in relaxation of vascular smooth muscle. Phosphodiesterase inhibitors exert their vasodilator effects through vascular smooth muscle relaxation resulting from increased levels of molecular second messengers ‐ monophosphate nucleotides. Prostacyclin, a prostaglandin member of the eicosanoid family, exerts vasodilator effects through smooth muscle relaxation by a similar mechanism as phosphodiesterase inhibitors, hence the theory of prostacyclin analogues. Thromboxane is another member of the eicosanoid family, but opposite to prostacyclin, it has a vasoconstrictive effect on vessels, leading to the theory of thromboxane synthase inhibitors.
Why it is important to do this review
Currently, limited treatment options are available for people with primary RP. Calcium channel blockers, the most commonly used class of drugs, have potential side effects including headache, postural hypotension, peripheral oedema, and constipation (Abernethy 1999). Furthermore, most studies have examined RP in the context of secondary conditions; therefore different treatments may have differing levels of effectiveness in primary compared to secondary RP.
This is the second update of the review first published in 2008 (Vinjar 2008). For this update, we have expanded the scope of the review to include any treatments given by any administration route including topical and intravenous vasodilators. We present all current evidence for vasodilatory drugs for treatment of primary RP, with the aim of guiding decision‐making among healthcare professionals and patients alike.
Objectives
To assess the effects of drugs with vasodilator effects on primary Raynaud's phenomenon as determined by frequency, severity, and duration of vasospastic attacks; quality of life; adverse events; and Raynaud Condition Score.
Methods
Criteria for considering studies for this review
Types of studies
We included all randomized controlled trials (RCTs) comparing a drug with vasodilator effects versus placebo for treatment of primary Raynaud's phenomenon (RP). We considered any method of randomization, and we included both parallel and cross‐over studies. We applied no language restrictions. Treatment with calcium channel blockers (CCBs) is addressed in another Cochrane systematic review (Ennis 2016).
Types of participants
We included trials involving participants with primary RP. We included trials with a mixture of primary and secondary RP if participants with primary RP could be identified and if data for this subgroup could be extracted. Study authors' definition of primary RP was accepted, unless details in the description of clinical characteristics of participants with primary RP did not comply with current diagnostic criteria or were deemed to be indicative of secondary RP or an alternative disease process.
Types of interventions
We included trials comparing oral, topical, or intravenous administration of any drug with vasodilator effects versus placebo or other drugs, including drugs registered for treatment of cardiovascular disease and genitourinary disease (Chapters 7 and 13 of the Australian Medicines Handbook [AMH] ‐ AMH 2018 ‐ and as Class C of the Anatomical Therapeutical Classification [ATC] system ‐ ATC classification).
Medications that were not defined as having vasodilatory properties, as contraindicated, or as recommended to be used with caution in patients with pre‐existing peripheral vascular disease or impaired circulation were excluded. This included all types of beta blockers and oxerutins. Other drugs with vasodilatory effects or with both vasodilatory and calcium channel blocking‐actions that have been suggested for treatment of RP but are primarily used for non‐cardiovascular diseases were included. See Table 1 for details of drugs with vasodilatory effects. We excluded all studies providing treatment with, or comparison with, calcium channel blockers or alternative (complementary) medicine or non‐pharmacologic modalities. We also excluded studies that failed to evaluate all required outcomes.
Owing to daily and seasonal variation in the frequency and duration of attacks, we excluded trials providing treatments administered only once (single‐dose trials) or over a period shorter than one week.
Types of outcome measures
Primary outcomes
-
Frequency of attacks
-
Severity of attacks or severity of symptoms during attacks measured on validated scales (e.g. pain and numbness on visual analogue scales or Likert scales; cold sensitivity as Cold Intolerance Severity Score [CISS])
-
Duration of attacks
-
Quality of life scores (measured by Health Assessment Questionnaires [HAQs], Short Form‐36 [SF‐36], or QuickDASH Outcome Measure)
-
Adverse events (including withdrawals)
Secondary outcomes
-
Raynaud Condition Score
-
Capillaroscopic flow/skin perfusion (measured by Doppler ultrasound or laser Doppler ultrasound imaging)
Search methods for identification of studies
We applied no language restrictions.
Electronic searches
The Cochrane Vascular Information Specialist first searched the following databases for relevant trials (July 31, 2014).
-
Cochrane Vascular Specialised Register (July 31, 2014).
-
Cochrane Central Register of Controlled Trials (CENTRAL; 2014, Issue 6), in the Cochrane Library, via the Cochrane Register of Studies Online.
See Appendix 1 for details of the search strategy used to search CENTRAL.
The Information Specialist also searched the following trial registries for details of ongoing and unpublished studies, using the term "raynaud" (July 31, 2014).
-
ClinicalTrials.gov (clinicaltrials.gov).
-
World Health Organization International Clinical Trials Registry Platform (who.int/trialsearch).
The Cochrane Vascular Information Specialist subsequently conducted systematic searches of the following databases for RCTs and controlled clinical trials without language, publication year, or publication status restrictions.
-
Cochrane Vascular Specialised Register via the Cochrane Register of Studies (CRS‐Web, searched on November 16, 2020).
-
Cochrane Central Register of Controlled Trials (CENTRAL), in the Cochrane Library, via Cochrane Register of Studies Online (CRSO 2020, issue 10).
-
MEDLINE (Ovid MEDLINE Epub Ahead of Print, In‐Process & Other Non‐Indexed Citations, Ovid MEDLINE Daily, and Ovid MEDLINE) (searched from July 31, 2014, to November 16, 2020).
-
Embase Ovid (searched from July 31, 2014, to November 16, 2020).
-
Cumulative Index to Nursing and Allied Health Literature (CINAHL) Ebsco (searched from July 31, 2014, to November 16, 2020).
-
Allied and Complementary Medicine Database (AMED) Ovid (searched from July 31, 2014, to November 16, 2020).
The Information Specialist modelled search strategies for other databases on the search strategy designed for CENTRAL. When appropriate, these strategies were combined with adaptations of the highly sensitive search strategy designed by Cochrane for identifying RCTs and controlled clinical trials (as described in the Cochrane Handbook for Systematic Reviews of Interventions, Chapter 6; Lefebvre 2011). Search strategies for major databases are provided in Appendix 2.
The Information Specialist searched the following trials registries on November 16, 2020.
-
World Health Organization International Clinical Trials Registry Platform (who.int/trialsearch).
-
ClinicalTrials.gov (clinicaltrials.gov).
Searching other resources
We did not check other resources.
Data collection and analysis
Selection of studies
All members of the review team identified trials for possible inclusion and independently reviewed abstracts and full‐text articles as appropriate. Abstracts were translated for the screening process and were included in the full‐text reviews if they appeared to fulfill selection criteria, or if they did not provide enough information to allow a decision regarding exclusion. We obtained full‐text articles and translated them as necessary via the Cochrane Task Exchange website (taskexchange.cochrane.org) or with the assistance of personal contacts who were native speakers. Translators were advised to follow inclusion criteria and to specify reasons for exclusion. If there was any uncertainty regarding the study, a full, direct translation would be requested, after which a member of the review team would review the article for inclusion. Two members of the team (KS and HA) contacted study authors for additional information to be used for inclusion or exclusion of studies. Two pharmaceutical companies and two study authors were contacted for additional information. All disagreements regarding article selection were discussed and were resolved between members, and if consensus could not be reached, KS or JN would review and make a final decision.
Data extraction and management
Each study was independently reviewed by two members of the review team (KS, MS, HA, JK). Data from included studies were extracted and recorded using Covidence (covidence.org), as recommended by Cochrane. Recorded data included sponsorship source, country where the study took place, type of setting (inpatient versus outpatient, hospital versus research center), author contact details, design of the study (randomization, blinding, and cross‐over versus parallel), baseline population characteristics, inclusion and exclusion criteria, types of interventions including dosing and route of administration, and finally, outcome measures of the study. We decided prior to data extraction that in the event that studies reported results at two different doses, results on the higher dose should be used in meta‐analyses. It is an assumption of meta‐analysis that studies are independent, that is, two or more results from a single study cannot be included in a meta‐analysis. We chose the result associated with the higher dose, assuming a dose‐effect relationship, as we are interested in identifying evidence of any effect.
Assessment of risk of bias in included studies
Two members of the author review team independently assessed the methodological quality of included trials (KS, MS, HA, EK, JK), using Cochrane's "Risk of bias" tool (Higgins 2011), and we recorded our judgements in Covidence. Domains assessed included sequence generation, allocation concealment, blinding of participants and personnel, blinding of outcome assessors, incomplete outcome data, selective outcome reporting, and any other sources of bias that the review team believed could affect the quality of a study. We judged each domain as being at low high or unclear risk of bias. We resolved disagreements through discussion.
Measures of treatment effect
We performed statistical analysis according to the statistical guidelines provided to review authors by Cochrane Vascular and described in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011). We carried out statistical analyses using the Web‐based Review Manager software (RevMan Web 2019).
For continuous data, we extracted mean differences (MDs) and their standard errors (SEs). SE incorporates both standard deviation (SD) and the number of subjects for the mean of the differences or the difference of the means and is the parameter that when estimated enables results from both parallel and cross‐over trials to be pooled. This is the main reason that we used SE over SD throughout the study. If different scales were used to measure effects, we planned to use standardized mean differences (SMDs) with 95% CIs when it was possible to meta‐analyze data from studies.
Frequency of attacks was reported variably and was converted to attacks per week when possible. Most studies reported mean and SD of frequency of attacks, and we converted these to MD (drug to placebo) and SE.
Duration of attacks was recorded in minutes. Most studies reported mean and SD of duration of attacks, and we converted these to MD (drug to placebo) and SE. Severity of attacks was assessed by a variety of methods. In each case, results were converted to MD (drug to placebo) and SE. Vayssairat 1996 utilised a severity measure on a 1 to 4 scale, and study authors did not state whether this was a categorical or a continuous measure but reported mean and SD; therefore we continued this reporting method. We used MD to report quality of life (QoL). As different VAS scales were used, we calculated SMD with 95% CI. We measured adverse events as risk ratios (RRs) with 95% CIs. Raynaud Condition Score was presented either out of 10 or out of 100. We converted all to a score out of 10 and then to MD (drug to placebo) and SE. Capillaroscopic flow/skin perfusion was reported by a range of measurement techniques for blood flow in the skin, and we calculated MD and SE between groups with 95% CI.
Four studies were parallel trials for which results were easily converted to the MD (SE) format (Chung 2009; Le Quentrec 1991; Rajagopalan 2003; Vayssairat 1996). Rustin 1987 provided results from a cross‐over trial already in the appropriate format (and corrected for period effect). Six studies were cross‐over trials that presented results so that it was possible to calculate the mean of the individual difference and SE (Caglayan 2012; Challenor 1991; Ettinger 1984; Madsen 1984; NCT01090492; Teh 1995). Van de Wal 1987 and Sovijarvi 1984 provided only group means for their cross‐over studies, thus preventing an accurate estimate of SE. In these cases, we used individual correlation between placebo and drug responses to estimate SE. We chose a conservative value of 0.5, as this was smaller than the correlations observed in the other studies (0.8 to 0.95) and provided the study with equivalent power to a parallel study with twice the number of participants. No changes from baseline values were included in the analyses. Two cross‐over studies presented dichotomous data that were reported narratively, as it was not possible to present these data accurately in an analysis format (Jaffe 1980; Nahir 1986).
Unit of analysis issues
All studies used the individual participant as the unit of analysis. All cross‐over studies were assessed to determine whether the washout period between study arms was sufficient to ensure no carrying over of effect between the first and second periods. Five studies did not include washout periods (Jaffe 1980; Madsen 1984; Rustin 1987; Sovijarvi 1984; Teh 1995). The medications used in these studies were oral captopril, topical glyceryl trinitrate, and moxisylyte, which have a half‐life (t1/2) of two hours, six minutes, and two hours, respectively. Even in the absence of the washout period, elimination time was less than or equal to 10 hours (based on five half lives, or the time for a medication to reach majority elimination (Ito 2011)), the drug effect dissipated by the next day. Challenor 1991 had three days of drug titration to minimize the side effect of enalapril, which has a t1/2 of 11 hours. It is safe to conclude that the drug effect was eliminated after or during the titration period (more than five t1/2). Van de Wal 1987 and NCT01090492 indicated two‐week washout periods, which was longer than the five t1/2 of ketanserin and PF‐00489791, respectively. Nahir 1986, Caglayan 2012, and Ettinger 1984 included one‐week washout periods, which were longer than the five t1/2 of glyceryl trinitrate. No analyses for cross‐over trials were carried out using results from only the first treatment period.
Two trials measured outcomes on more than one occasion (Le Quentrec 1991; Vayssairat 1996). Le Quentrec 1991 measured outcomes at two, four, and six months, using the same dose during the whole treatment period. Differences in outcomes between four and six months were minor. We used the six‐month outcomes in this review. Vayssairat 1996 measured outcomes after phase 2, 20 μg three times daily, and phase 3, 40 μg three times daily. We used the latter outcomes in this review under the assumption of a dose‐response relationship.
We used the generic inverse variance (GIV) option in RevMan Web 2019 so that results from cross‐over trials could be combined with results from parallel trials. Standard errors (SEs) had to be estimated from the cross‐over trials and could then be entered along with main effects (as mean differences) using the GIV option and combined with SEs and MDs from parallel studies. No information was provided for active and placebo groups individually, nor for numbers in each treatment group, as these do not make sense specifically in the context of cross‐over trials.
Dealing with missing data
One member of the review team (KS or HA) contacted study authors or trial coordinators (one was a pharmaceutical company) for additional information when data were believed to be missing (Caglayan 2012; Chung 2009; NCT01090492; Rajagopalan 2003; Vayssairat 1996). Chung 2009 included an intention‐to‐treat discussion that included all randomized participants who had applied at least one dose of the study drug and had recorded data in their electronic diaries. These investigators utilized only observed data and performed no imputation. Caglayan 2012 performed an intention‐to‐treat analysis. Additional information for Vayssairat 1996 clarified that intention‐to‐treat analyses had been carried out. For all other trials with exclusions, withdrawals, or losses to follow up, no intention‐to‐treat analysis was discussed or presented (Nahir 1986; NCT01090492; Rajagopalan 2003; Sovijarvi 1984; Teh 1995). In these, calculations of results for the cross‐over trials had been based on participants who completed both treatment arms. No data were re‐analyzed according to the principles of intention‐to‐treat.
Assessment of heterogeneity
We used RevMan Web for meta‐analysis (RevMan Web 2019), and the heterogeneity statistics presented consist of Chi² and P value for Cochrane's heterogeneity statistic Q, along with I², yielding a percentage of total variation across studies due to heterogeneity above what is expected by chance. A value of 0% indicates no observed heterogeneity, and larger values show increasing heterogeneity. Analyses with I² > 50% and P(Q) < 0.05 were performed with a random‐effects model (traditional Der Simonian and Laird method) (DerSimonian 1986). A random‐effects model was also used if there was empirical evidence for heterogeneity; otherwise a fixed‐effect model was used. Tau², the estimated variance of effect under a random‐effects analysis, was also presented in this instance.
Assessment of reporting biases
We were not able to construct a funnel plot to assess the possibility of publication bias because of lack of comparable studies.
Data synthesis
We carried out meta‐analyses using RevMan Web and guidelines provided by Cochrane Vascular (RevMan Web 2019). We performed statistical analysis according to the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011). As both parallel and cross‐over trials were included in this meta‐analysis, we used the methods described in Curtain 2002 and Elbourne 2002 to appropriately combine results from all eligible studies. We carried out meta‐analyses using MD (and SE) between treatment and control post‐intervention values. In the case of parallel trials, this was the difference in the means of treatment and placebo groups. For cross‐over trials, this was the mean of the differences for each individual between treatment and placebo phases. We performed meta‐analyses using weighted mean differences (MDs) when the unit of measure was the same or was very similar (e.g. frequency of attacks), or using SMDs where there was heterogeneity in measurement methods (e.g. severity or capillaroscopic flow/skin perfusion). We presented results for drugs separately because of lack of pharmacological similarity.
Subgroup analysis and investigation of heterogeneity
We were unable to perform subgroup analyses involving gender, age, and duration of attacks due to insufficient data regarding separate participant groups amongst the included studies. We were unable to determine numbers needed to treat due to limited results. Potential analyses that could prove useful in establishing treatment efficacy in the future, should new studies be included, involve gender, age, and duration of attack subgroups. We were, however, able to group and present the data across all included studies by pharmacological class. This allowed for comparisons and assessments of qualitative and quantitative interactions between different classes of vasodilators. Pre‐existing subgroup analyses from individual studies were included and discussed if relevant to the outcomes of this meta‐analysis.
Sensitivity analysis
Challenor 1991 received very high weight for frequency of attacks compared to the other studies because of a remarkably small reported SD of frequency of attacks for both groups. Because of this, we performed a sensitivity analysis excluding the enalapril trial from the meta‐analysis. Due to the limited number of available studies for each comparison, we were unable to carry out any further sensitivity analyses.
Summary of findings and assessment of the certainty of the evidence
We created "Summary of findings" tables according to Higgins 2011 and used the GRADE method to present the most important findings of this systematic review (Atkins 2004). We used GRADEproGDT software to assist in preparation of "Summary of findings" tables (GRADEpro GDT 2015). We included the six outcomes of greatest clinical relevance to patients and healthcare professionals: frequency of attacks, duration of attacks, severity of attacks, QoL, adverse events, and RCS. We created one table for each drug type: ACE inhibitor compared to placebo for primary Raynaud's phenomenon (summary of findings Table 1); alpha blockers compared to placebo for primary Raynaud's phenomenon (summary of findings Table 2); prostaglandin/prostacyclin analogues compared to placebo for primary Raynaud's phenomenon (summary of findings Table 3); thromboxane synthase inhibitors compared to placebo for primary Raynaud's phenomenon (summary of findings Table 4); selective serotonin reuptake inhibitors compared to placebo for primary Raynaud's phenomenon (summary of findings Table 5); nitrate/nitrate derivatives compared to placebo for primary Raynaud's phenomenon (summary of findings Table 6); and phosphodiesterase inhibitors compared to placebo for primary Raynaud's phenomenon (summary of findings Table 7); The certainty of evidence for each outcome was determined by the GRADE approach, which considers overall risk of bias of included studies, directness of evidence, inconsistency within results, precision of estimates, and risk of publication bias (Atkins 2004).
Results
Description of studies
See Figure 1.
Study flow diagram.
Results of the search
We identified seven new studies for this update (Caglayan 2012; Chung 2009; Nahir 1986; NCT01090492; Rajagopalan 2003; Sovijarvi 1984; Teh 1995); excluded 22 new studies (Allegra 1983; Arcas Meca 1972; Barry 2000; Belch 1985; Bellucci 1987; Belluci 1990; Brotzu 1989; Clement 1980; Diehm 1983; Fischer 1985; JaniniDa 1988; Jenkins 2013; Kingma 1995; Kirichenko 1991; Lee 2014; McFadyen 1973; Mirza 2019; NCT00419419; NCT0048776; Roustit 2017; Strozzi 1982; Surwit 1982); identified two ongoing studies (EUCTR2005‐000295‐41‐DE; NCT02583789), and assessed one study as awaiting classification (Sakaguchi 1990).
Included studies
See the Characteristics of included studies table.
We identified seven new studies for this update (Caglayan 2012; Chung 2009; Nahir 1986; NCT01090492; Rajagopalan 2003; Sovijarvi 1984; Teh 1995). Two studies were published and made available since the last update (Caglayan 2012; NCT01090492), and four were added due to the change in our study selection criteria to expand from oral vasodilators to any vasodilator (Chung 2009; Nahir 1986, Sovijarvi 1984; Teh 1995). The final additional study was excluded by the previous review authors on the basis of not reporting numbers of participants in treatment and control groups (Rajagopalan 2003). This brings the total to 15 included studies with 635 participants (Caglayan 2012; Challenor 1991; Chung 2009; Ettinger 1984; Jaffe 1980; Le Quentrec 1991; Madsen 1984; Nahir 1986; NCT01090492; Rajagopalan 2003; Rustin 1987; Sovijarvi 1984; Teh 1995; Van de Wal 1987; Vayssairat 1996). The largest included study was Chung 2009, which included 140 participants. The remainder of the studies involved numbers of participants ranging from 6 in Ettinger 1984 to 125 in Vayssairat 1996.
Of the seven new studies included in this update, one was referenced as an ongoing study (NCT01090492), and four investigated topical formulations and therefore were not within the scope of the previous review (Chung 2009; Nahir 1986; Sovijarvi 1984; Teh 1995). One was published in August of 2012 after publication of the previous version of the review (Caglayan 2012). The remaining study was excluded by the previous review authors on the basis of not reporting numbers of participants in treatment and control groups (Rajagopalan 2003). Inclusion in this review was allowed, as we were able to calculate the number of participants in each group based on means and standard deviations provided in the original study.
Variation in the included studies was notable in terms of included participant characteristics and outcome measures. A consensus definition for primary RP was established only in 2014 (Maverakis 2014), with four different criteria described before this (Maverakis 2014). Additionally, there were distinct differences in terms of important baseline demographics such as gender, smoking, and weather and setting of the study; furthermore, baseline disease severity demonstrated variation based on various outcome measures.
The included studies were published between 1989 and 2013. A total of 11 medications with vasodilating effects were represented. The included studies used oral and non‐oral routes of administration and topical therapy. We excluded studies that mentioned intravenous vasodilators, as we excluded trials with treatments administered only once.
Participants
Five studies included participants with both primary and secondary Raynaud's phenomenon (RP) (Caglayan 2012; Chung 2009; Nahir 1986; NCT01090492; Rajagopalan 2003). Jaffe 1980 included a combination of primary RP and chilblains but separately presented results for each group and therefore was included. All remaining studies included only patients with primary RP (Challenor 1991; Ettinger 1984; Jaffe 1980; Le Quentrec 1991; Madsen 1984; Rustin 1987; Sovijarvi 1984; Teh 1995; Van de Wal 1987; Vayssairat 1996).
All studies defined primary and secondary RP using varying definitions. Nahir 1986 did not discuss its definition of primary RP. Le Quentrec 1991 defined primary disease as diagnosed by "classical criteria," with no references to define this. Challenor 1991 defined primary RP using the Blunt and Porter 1981 definition. Caglayan 2012 included a link for inclusion and exclusion criteria, but at the time of writing this review, the link was defunct. Heterogeneity was significant in the definition of secondary RP, with the general theme being the absence of any biochemical, serological, or clinical findings including digital ulceration.
The baseline frequency of attacks was similar to that in previous reviews, at 7 to 21 per week in all trials except Van de Wal 1987, which reported a mean of 91 attacks per week. It is unlikely that this is an error in reporting. Overall duration of disease was between 1 and 50 years, with an approximate median of 15 to 20 years. Ten of 41 participants in this trial had previously undergone thoracic sympathectomy, reflecting severe disease.
One study included only female participants (Rustin 1987); others included more female than male participants, reflecting the prevalence of RP in the population. Median age of participants was generally consistent, at between 20 and 40 years, with an overall range of 21 to 74 years.
The studies took place in different settings, including a laboratory setting for Sovijarvi 1984; a general practice for Jaffe 1980; and specialist outpatient clinics for seven studies (Caglayan 2012; Chung 2009; Ettinger 1984; Le Quentrec 1991; Rajagopalan 2003; Rustin 1987; Vayssairat 1996). Two multicenter studies included a combination of universities and hospital clinical settings (NCT01090492; Vayssairat 1996). The setting was not described in five trials (Challenor 1991; Madsen 1984; Nahir 1986; Teh 1995; Van de Wal 1987).
All trials were conducted during the winter, and one was conducted from winter to early spring (Chung 2009). The prevalence of smoking was reported by most study authors and varied between 5% and 53%. Two studies included active smoking as an exclusion criterion (NCT01090492; Rajagopalan 2003). Studies differed in their details and descriptions of exclusion criteria.
Interventions
See additional Table 2 and Table 3.
| AMH Chapter | Class | ATC Code | Trials |
|---|---|---|---|
| Cardiovascular drugs | Vasodilators, antihypertensive drugs, and prostacyclin analogues | C02D, | Included: Rajagopalan 2003; Vayssairat 1996 Ongoing: NCT02583789 Awaiting classification: Sakaguchi 1990 |
| Cardiovascular drugs | Phosphodiesterase type 5 inhibitors | G04BE | Included: Caglayan 2012; NCT01090492 Ongoing: EUCTR2005‐000295‐41‐DE |
| Cardiovascular drugs/endocrine drugs | Alpha adrenoreceptor‐blocking drugs | C02C | Included: none |
| Unlisted | Serotonin antagonists | C02KD | Included: Van de Wal 1987 |
| Cardiovascular drugs | Angiotensin‐converting enzyme inhibitors | C09A | Included: Challenor 1991; Madsen 1984; |
| Cardiovascular drugs | Angiotensin‐II receptor antagonists | C09C | Included: none |
| Cardiovascular drugs | Nitrates | C01D | Included: Chung 2009; Nahir 1986; Teh 1995 |
| Unlisted | Peripheral vasodilators and related drugs | C04A | Included: Jaffe 1980; Le Quentrec 1991 |
| Psychotropic drugs | Selective serotonin reuptake inhibitors | N06A B | Included: none |
| Allergy and anaphylaxis | Antihistamines | N07CA | Included: none |
| Other (including Bradilan, Dazoxiben, UK‐38, 485 (dazmegrel)) | Included: Ettinger 1984 |
AMH: Australian Medicines Handbook.
ATC: Anatomical Therapeutical Classification.
aComparison between udenafil and amlodipine.
bThis study mentions ketanserin and pentoxifylline.
cThis study mentions alpha‐methyldopa, guanethidine, and debrisoquine.
| AMH | Class | ATC code | Drug name |
|---|---|---|---|
| Cardiovascular drugs | Vasodilators, antihypertensive drugs including | C02D, B01A | Included: beraprost sodium, cilostazol |
| Cardiovascular drugs | Phosphodiesterase inhibitors/phosphodiesterase type 5 inhibitors | G04BE | Included: vardenafil Ongoing: vardenafil |
| Cardiovascular drugs/endocrine drugs | Alpha adrenoreceptor‐blocking drugs | C02C | Included: thymoxamine |
| Cardiovascular drugs | Angiotensin‐converting enzyme inhibitors | C09A | Included: captopril, enalapril |
| Cardiovascular drugs | Angiotensin‐II receptor antagonists | C09C | Included: none |
| Cardiovascular drugs | Nitrates | C01D | Included: nitroglycerin/glyceryl trinitrate |
| Unlisted | Peripheral vasodilators and related drugs | C04A | Included: buflomedil, thymoxamine |
| Psychotropic drugs | Selective serotonin reuptake inhibitors | N06AB | Included: ketanserin |
| Allergy and anaphylaxis | Antihistamines | N07CA | Included: none |
| Others | Included: dazoxiben |
AMH: Australian Medicines Handbook.
ATC: Anatomical Therapeutical Classification.
Among the studies, seven classes of drugs that cause vasodilation were represented: ACE inhibitors, selective serotonin reuptake inhibitors, nitrates/nitrate derivatives, phosphodiesterase‐5 inhibitors, prostacyclin/prostanoids, alpha blockers, and thromboxane synthase inhibitors. Table 1 presents classifications of drugs with vasodilator effects from the Australian Medicines Handbook. Thromboxane synthase inhibitors were included in this study, although they are not included in Table 1. Two trials involved captopril (Madsen 1984; Rustin 1987), and four involved glyceryl trinitrate (GTN) (Chung 2009; Nahir 1986; Sovijarvi 1984; Teh 1995), one of which involved MQX‐503, a novel vehicle for delivery (Chung 2009). The rest were single trials on each of the following: beraprost (Vayssairat 1996), buflomedil (Le Quentrec 1991), cilostazol (Rajagopalan 2003), dazoxiben (Ettinger 1984), enalapril (Challenor 1991), ketanserin (Van de Wal 1987), PF‐00489791 (NCT01090492), moxisylyte (Jaffe 1980), and vardenafil (Caglayan 2012).
Twelve studies compared the drug with a single placebo (Caglayan 2012; Challenor 1991; Chung 2009; Jaffe 1980; Le Quentrec 1991; Madsen 1984; Nahir 1986; NCT01090492; Rajagopalan 2003; Rustin 1987; Van de Wal 1987; Vayssairat 1996). Sovijarvi 1984 compared GTN against two sets of placebo. Ettinger 1984 compared dazoxiben and a calcium channel blocker as single and combined agents against placebo. Calcium channel blocker data have been excluded from this review. Five studies tested topical agents (Chung 2009; NCT01090492; Nahir 1986; Sovijarvi 1984; Teh 1995), and the remaining 10 assessed oral medications (Caglayan 2012; Challenor 1991; Ettinger 1984; Jaffe 1980; Le Quentrec 1991; Madsen 1984; Rajagopalan 2003; Rustin 1987; Van de Wal 1987; Vayssairat 1996)
Length of studies
Lengths of treatment in studies varied from weeks to months, with the longest lasting six months (Le Quentrec 1991). Teh 1995 and Sovijarvi 1984 had a one‐week run‐in period before the start of the study, Van de Wal 1987 had a run‐in period of four weeks. Four studies had two‐week run‐in periods (Chung 2009; Ettinger 1984; Rustin 1987; Vayssairat 1996). Four studies had no run‐in period (Challenor 1991; Jaffe 1980; Le Quentrec 1991; Madsen 1984). For some cross‐over studies, one‐week washout periods were utilized (Caglayan 2012; Ettinger 1984; Nahir 1986). All studies were double‐blind during the active treatment phase, except Chung 2009, which was triple‐blinded according to obtained trial data. Caglayan 2012 had an extra four‐week follow‐up period, Chung 2009 had a one‐week follow‐up, and Van de Wal 1987 had a 12‐week long‐term follow‐up.
Outcomes
Subjective measurements were reported by all studies except Sovijarvi 1984. These measures included documentation by participants regarding frequency of attack in Madsen 1984 and dual reporting of frequency and severity in other studies (Challenor 1991; Jaffe 1980; Le Quentrec 1991; Nahir 1986; Rajagopalan 2003; Teh 1995; Vayssairat 1996). Visual analogue scale (VAS) scores were also reported (Challenor 1991), as was the Raynaud Condition Score (RCS) (Caglayan 2012; Chung 2009). Despite measuring the same outcomes, scales used between different studies varied substantially. Number of attacks was measured as episodes per day (Challenor 1991; Le Quentrec 1991), as daily mean (Madsen 1984), as the mean of attacks at completion of the study (Rajagopalan 2003), or simply as dichotomous outcomes (Jaffe 1980; Nahir 1986). Severity of attacks lacked any standardization between chosen numerical scales (Challenor 1991; Le Quentrec 1991; Rajagopalan 2003; Teh 1995, Vayssairat 1996), or dichotomous outcomes were reported (Nahir 1986). Duration of attacks was measured by Chung 2009Jaffe 1980NCT01090492Rustin 1987, and Van de Wal 1987. Teh 1995 was the only study that provided a quality of life score.
Objective measures used included Doppler evaluation/capillaroscopy rheographic evaluation of blood flow (Sovijarvi 1984). Rajagopalan 2003 and Rustin 1987 used laser Doppler flowmetry to measure blood flow. Rustin 1987 used digit II and expressed the result in volts, Rajagopalan 2003 used digit III and expressed flow as the peak perfusion ratio. Sovijarvi 1984 and Van de Wal 1987 measured digital systolic blood pressure (mmHg) by venous occlusion plethysmography of digit II. Van de Wal 1987 also used digit IV (we report results for digit II). Adverse events were recorded in all trials.
Excluded studies
We excluded 22 new studies for this update (Allegra 1983; Arcas Meca 1972; Barry 2000; Belch 1985; Bellucci 1987; Belluci 1990; Brotzu 1989; Clement 1980; Diehm 1983; Fischer 1985; JaniniDa 1988; Jenkins 2013; Kingma 1995; Kirichenko 1991; Lee 2014; McFadyen 1973; Mirza 2019; NCT00419419; NCT0048776; Roustit 2017; Strozzi 1982; Surwit 1982). In total, we excluded 59 full‐text studies. See Characteristics of excluded studies.
Six studies were excluded because participants did not meet the inclusion criteria, mainly due to a prior diagnosis of secondary Raynaud's phenomenon (Belch 1995; Kirichenko 1991; Lee 2014; Marasini 2004; Surwit 1982; Torley 1990). One study was excluded as the population group was not clearly defined (Clement 1986). Five studies were excluded as a single dose of medication was administered only once (Courbier 1981; McFadyen 1973; Mohrland 1985; Tucker 1999; Wesseling 1981). Two studies were excluded for use of a single dose of treatment (Seibold 1986; Shawket 1991). Five studies described intervention or medications that was not within the scope of this review, such as use of calcium channel blockers, and were excluded (Allegra 1983; Brotzu 1989; Dumoulin 1981; Dziadzio 1999b; Sunderland 1988). Many of the studies that were excluded included fewer than 10 participants with primary Raynaud's phenomenon. Moreover, multiple studies also had more than one reason for exclusion.
Thirty‐two studies were excluded because we were unable to extract the data specific for primary Raynaud's participants (Arcas Meca 1972; Arosio 1989; Bali 2011; Barry 2000; Belch 1983; Belch 1985; Bellucci 1987; Clement 1980; Coffman 1989; Coleiro 2001b; Courbier 1981; Davinroy 1993; Diehm 1983; Fischer 1985; Friedman 2007; JaniniDa 1988; Jenkins 2013; Kahan 1985; Kingma 1995; Kyle 1992; Longstaff 1985; Luderer 1984; Marasini 1988; Maurel 1995; Mirza 2019; NCT00419419; Nilsen 1979; Roustit 2017; Russell 1985; Strozzi 1982; Tooke 1990; Wollersheim 1986). This was mainly due to presentation of combined results for all participants and we were not able to clearly differentiate between participants with primary and secondary Raynaud's phenomenon. Grigg 1989a was excluded for including an outcome measurement that was not within the defined primary or secondary outcomes.
In the previous version of the review, Cleophas 1984 was excluded because some participants with primary Raynaud's tested positive for anti‐nuclear antibodies (ANAs). We revised the reason for exclusion for Cleophas 1984 to the use of a beta blocker as a comparator, lack of a placebo group, and no subgroup analysis.
The following classes of drugs or single drugs were not represented in the included studies as the study design did not meet inclusion criteria: angiotensin‐II receptor antagonists, antihistamines, botulinum toxin, bradilan, CL115,347 (Cyanamid International), debrisoquine, guanethidine, pentoxifylline, reserpine, and suloctidil (see additional Table 2 and Table 3). Some medications are not mentioned in Table 1, as this is not an exhaustive list of medications with any vasodilatory effect. This was done to ensure that the results discussed and analyzed were specific to vasodilators.
Two new ongoing studies were identified for this update (EUCTR2005‐000295‐41‐DE; NCT02583789; see Ongoing studies). One study was assessed as awaiting classification (Sakaguchi 1990; see Studies awaiting classification).
Risk of bias in included studies
See Figure 2 and Figure 3 for details on risk of bias.
Risk of bias graph: review authors' judgements about each risk of bias item presented as percentages across all included studies.
Risk of bias summary: review authors' judgements about each risk of bias item for each included study.
Allocation
Four of the 15 studies included in this review reported their random sequence generation methods appropriately and were found to have low risk (Chung 2009; Le Quentrec 1991; Rustin 1987; Vayssairat 1996). One study utilized a random number system (Rustin 1987), and three used a randomization table (Chung 2009; Le Quentrec 1991; Vayssairat 1996). The remainder of the studies had unclear risk of bias regarding how their sequences were generated, as they failed to disclose their methods of randomization. The supplementary data supplied by Caglayan 2012 described a form of constrained randomization utilizing permuted blocks of lengths four and six that were computer‐generated, sex‐stratified and in pseudo‐random sequence; given the binary nature of the stratification, this study was deemed to be at high risk of bias. With regard to allocation concealment; one of the 15 studies clearly and adequately described its methods and was deemed to be at low risk of bias (Le Quentrec 1991). Chung 2009 did not publish its study methods but instead included them publicly online in NCT00266669 2005 risk of bias was assessed based on this information. Nahir 1986 included a purely female population, resulting in high risk of bias. For the remainder of the included studies, no information was stated or could be elicited regarding how allocation was concealed. This lack of disclosure led to the remainder of studies holding unclear risk with regard to bias.
Blinding
Performance bias
Regarding blinding of participants and personnel, four studies were deemed to have low risk of bias because they were blinded with identical or matching placebo (Ettinger 1984; Jaffe 1980; Le Quentrec 1991; Vayssairat 1996). Other studies failed to specify how double‐blinding was achieved, risking interviewer bias (Challenor 1991; Chung 2009; Caglayan 2012; Madsen 1984; NCT01090492; Nahir 1986; Rajagopalan 2003; Rustin 1987; Sovijarvi 1984; Teh 1995; Van de Wal 1987), resulting in a risk rating of "unclear." Madsen 1984 further failed to mention the frequency of evaluation with its interviewers.
Detection bias
One study appropriately identified the measures through which their outcome assessment remained blinded (Vayssairat 1996). The remaining studies were deemed to have unclear risk for the following reasons. Self‐reported outcome measures performed by the participants of six studies were used to collate data (Chung 2009; Ettinger 1984; Jaffe 1980; Nahir 1986; Vayssairat 1996; Teh 1995). Self‐reporting carries unavoidable bias given its subjective nature. Some studies attempted to control this risk of bias by asking participants to record quantitative measures such as number of attacks and by setting the precedent for these recordings to occur daily (Challenor 1991; Madsen 1984; Rustin 1987). However, it cannot be clarified whether all participants accurately recorded the number of attacks regularly, without consideration of other influencing variables that may have compounded on their compliance with instructions. Similarly, it was found that within Jaffe 1980, many participants were unable to quantify the severity of their symptoms appropriately and accurately.
Most of the included studies failed to disclose, if mentioned, how outcomes and statisticians were blinded, especially with respect to studies that utilized physical examination at the screening and completion phases of treatment (Caglayan 2012; Challenor 1991; Madsen 1984; NCT01090492; Rajagopalan 2003; Rustin 1987; Sovijarvi 1984; Teh 1995; Van de Wal 1987). Consequently, we were unable to confidently assess the level of bias they presented ‐ leaving this as unclear. Ettinger 1984Jaffe 1980Le Quentrec 1991, and Nahir 1986 did not indicate that their assessors were blinded, and so were assessed as unclear.
Incomplete outcome data
Eight included studies were at low risk of attrition bias, with all participants accounted for in the results and the significance of dropouts addressed (Challenor 1991; Chung 2009; Ettinger 1984; Le Quentrec 1991; Madsen 1984; Sovijarvi 1984; Van de Wal 1987; Vayssairat 1996). Two studies showed unclear risk of bias (Rajagopalan 2003; Rustin 1987), with unclear significance of dropouts for data projection and unclear formulation of results. Caglayan 2012 did not present dropout rate, although serious adverse events were noted, which suggested unclear risk of bias. NCT01090492 did not specify any incomplete outcome data within the online published study, also resulting in unclear attrition bias. The risk of attrition bias was considered high in Jaffe 1980, as five participants were excluded from the analysis for objectively atypical reasons ‐ two for not having experienced attacks of Raynaud's phenomenon during the trial, and a further three for experiencing adverse events whilst on placebo. Teh 1995 failed to provide discrete numerical values within its results section, resulting in analytical inadequacies throughout the presented data. Of 21 participants, two were excluded following incomplete data collection, and two others following unspecified adverse events related to the GTN. Although participant withdrawals were discussed, these participants were not included in the statistical analysis, and there was no discussion regarding the impact the withdrawals may have had on study data. We judged this study to be at high risk of bias. Nahir 1986 also had high risk of attrition bias, as there were three withdrawals from a comparatively small study size. Two participants withdrew from the treatment group due to headaches/dizziness, and one participant with recurrent dermatitis withdrew from an unspecified group.
Selective reporting
Eleven studies accounted for all pre‐specified outcome measures in their results, resulting in assignment of low risk of selective reporting bias (Challenor 1991; Chung 2009; Ettinger 1984; Jaffe 1980; Le Quentrec 1991; Madsen 1984; Nahir 1986; Rustin 1987; Sovijarvi 1984; Van de Wal 1987; Vayssairat 1996). NCT01090492 and Rajagopalan 2003 presented unclear risk with regard to reporting bias, as they failed to specify details regarding participant dropout and completion of data collection. Furthermore, Rajagopalan 2003 failed to define the number of participants in the control and intervention groups when reporting outcomes. However, based on means and standard deviations, we calculated the number of participants in each group and thus included this study in the meta‐analysis. Caglayan 2012 was deemed at high risk of subgroup analysis for performance of analysis only for its RCS group and not for the digital blood flow subsection. The website used to present study inclusion criteria was unavailable for checking the variable and invariable characteristics of participants against the final data. Teh 1995 was also deemed at high risk of reporting bias, as the means by which outcomes were reported varied throughout the results, thus allowing for bias through data reporting. Furthermore, participants were excluded from aspects of results formulation if their diary entries were considered "poor." Le Quentrec 1991 was rated as low risk, as all pre‐specified outcome measures were accounted for in the results, despite lack of clarification of criteria.
Other potential sources of bias
Eleven studies were at low risk of other potential sources of bias (Caglayan 2012; Chung 2009; Ettinger 1984; Le Quentrec 1991; Nahir 1986; NCT01090492; Rajagopalan 2003; Rustin 1987; Sovijarvi 1984; Teh 1995; Van de Wal 1987). Three studies were assessed as being at unclear risk of other bias because they were cross‐over studies, and although it is unlikely that the drugs had carry‐over effects, information detailing effects or how appropriate washout times were achieved was lacking (Challenor 1991; Jaffe 1980; Madsen 1984). Vayssairat 1996 was assessed as being at unclear risk, as study groups did not meet the size calculation.
Effects of interventions
See: Summary of findings 1 Angiotensin‐converting enzyme inhibitors compared to placebo for primary Raynaud's phenomenon; Summary of findings 2 Alpha blockers compared to placebo for primary Raynaud's phenomenon; Summary of findings 3 Prostagladin/prostacyclin analogues compared to placebo for primary Raynaud's phenomenon; Summary of findings 4 Thromboxane synthase inhibitors compared to placebo for primary Raynaud's phenomenon; Summary of findings 5 Selective serotonin reuptake inhibitors compared to placebo for primary Raynaud's phenomenon; Summary of findings 6 Nitrate/nitrate derivatives compared to placebo for primary Raynaud's phenomenon; Summary of findings 7 Phosphodiesterase inhibitors compared to placebo for primary Raynaud's phenomenon
The 15 included studies reported on 11 different vasodilators, using oral, non‐oral, and topical administration. We were not always able to carry out meta‐analysis due to limited data or to clinical and methodological differences. We have presented the results by drug class and by drug separately below.
Angiotensin‐converting enzyme (ACE) inhibitors
(Captopril [ATC code C09AA01] and enalapril [ATC code C09AA02], ACE inhibitors, plain)
Three cross‐over studies including two studies comparing captopril 25 mg three times daily ‐ Madsen 1984; Rustin 1987 ‐ and enalapril 20 mg once daily ‐ Challenor 1991 ‐ with placebo were included. All of the 46 included participants had primary Raynaud's phenomenon (RP). The duration for all captopril trials was two periods of six weeks and for the enalapril trial two periods of four weeks. See summary of findings Table 1.
Frequency of attacks
For the meta‐analysis of the ACE inhibitors captopril and enalapril, frequency of attacks may be increased with treatment (mean difference [MD] 0.79, 95% confidence interval [CI] 0.43 to 1.16; 44 participants, 3 studies; low‐certainty evidence; see Analysis 1.1).
Subgroup analysis revealed no clear differences in frequency of attacks between captopril and placebo groups (MD 0.72, 95% CI ‐1.73 to 3.16; 2 studies; low‐certainty evidence). Enalapril alone was associated with a small increase in the frequency of attacks per week (Challenor 1991). The difference in the mean number of attacks per week was 0.80 (95% CI 0.43 to 1.17; 20 participants, 1 study; low‐certainty evidence; see Analysis 1.1), favouring placebo. Thirteen of 20 participants experienced more attacks on enalapril than on placebo. No differences between subgroups was detected by the subgroup test (P = 0.94). A random‐effects model was used because captopril and enalapril belong to the same class of drugs but have different pharmacological properties, and the duration of treatment periods was unequal (DerSimonian 1986). Although there was similarity among participants in the two captopril trials, as well as in dose, length of treatment, and lack of a washout period, heterogeneity was high (Madsen 1984; Rustin 1987).
Severity of attacks
Challenor 1991 and Rustin 1987 reported on severity. No evidence suggested a difference in the severity of attacks between ACE inhibitors and placebo (MD ‐0.17, 95% CI ‐4.66 to 4.31; 34 participants, 2 studies; low‐certainty evidence; see Analysis 1.2). No differences between subgroups were detected with the subgroup test (P = 0.69).
Duration of attacks
Only Rustin 1987 reported on this; captopril did not appear to make any clear difference in duration of attacks (MD 0.54, 95% CI ‐1.34 to 2.42; 14 participants, 1 study; low‐certainty evidence; see Analysis 1.3). Rustin 1987 did not stipulate units of attack duration, but "minutes" seemed most likely and was assumed.
Capillaroscopic flow/skin perfusion
Only Rustin 1987 reported on this, describing no clear changes in capillaroscopic flow following treatment (MD 0.19, 95% CI ‐0.05 to 0.42; 13 participants, 1 study; low‐certainty evidence; see Analysis 1.4). These findings were imprecise.
Subjective assessment of improvement
Only Challenor 1991 reported on this outcome. When enalapril treatment was compared with placebo, an increased subjective assessment of improvement may be detected in the placebo group (10‐cm visual analogue scale) (MD 1.10, 95% CI 0.32 to 1.88; 1 study; low‐certainty evidence; see Analysis 1.5).
Adverse events
One individual from the Rustin 1987 trial, which featured captopril, withdrew due to pregnancy, and one participant withdrew from the Challenor 1991 trial, which involved enalapril. The withdrawal in the enalapril study was done for personal reasons and was deemed unrelated to treatment by study authors. No side effects were reported in the Rustin 1987 trial. Madsen 1984 reported one participant with nausea the first week and one with pain in the calf muscle during the whole intervention period in the captopril group, and no side effects in the placebo group. In the Challenor 1991 trial, nine participants receiving enalapril and eight participants receiving placebo reported side effects. Dizziness was most commonly reported, but all side effects were transient. Pooling of data revealed no clear evidence of a difference in adverse events between treatment and placebo groups (risk ratio [RR] 1.35, 95% CI 0.67 to 2.73; 46 participants, 3 studies; low‐certainty evidence; see Analysis 1.6). No differences between subgroups were detected with the subgroup test (P = 0.33).
Raynaud Condition Score
No study reported on the Raynaud Condition Score.
Alpha blockers
(Buflomedil [ATC code C04AX20] and moxisylyte [thymoxamine] [ATC code C04AX10])
Two studies used alpha blockers. One study included in this update investigated buflomedil (Le Quentrec 1991). It involved 31 participants with primary RP and compared buflomedil 300 mg twice daily with placebo in a parallel study design. We used outcomes measured at six months; this study reported outcomes also at two months and at four months. The second study compared moxisylyte 40 mg four times daily with placebo (Jaffe 1980). This study initially included 41 participants, all with primary RP, and 33 individuals completed the study. Participants did not complete the study for various reasons. This study had a cross‐over design; each treatment period lasted two weeks. See summary of findings Table 2.
Frequency of attacks
We were not able to combine the data in a meta‐analysis, as data were presented as number of attacks (dichotomous) in Jaffe 1980 and as difference in frequency in Le Quentrec 1991 (continuous).
In Le Quentrec 1991, the frequency of attacks per week was probably reduced with buflomedil treatment (MD ‐8.82, 95% CI ‐11.04 to ‐6.60; 31 participants, 1 study; moderate‐certainty evidence; see Analysis 2.1). Baseline frequency was 24 attacks per week in both intervention groups.
In Jaffe 1980, 33 participants with primary RP completed both treatment arms for the outcome "frequency of attacks." Nineteen participants had fewer attacks during the moxisylyte period, and 10 during the placebo period. Four participants had an equal number of attacks in each period. More participants on moxisylyte experienced less frequent attacks (P < 0.02; Wilcoxon matched pair signed rank test), as reported by study authors.
Severity of attacks
Both studies reported on severity of attacks, but we were unable to carry out a meta‐analysis, as they used different methods to present the data.
Le Quentrec 1991 results indicate that the severity of attacks was slightly reduced with buflomedil treatment (MD ‐0.41, 95% CI ‐0.52 to ‐0.30; 31 participants, 1 study; moderate‐certainty evidence; see Analysis 2.2).
In Jaffe 1980, only 25 participants were able to quantify the severity of the attacks ‐ study authors reported that not all participants were able to quantify the severity of attacks in a meaningful way. Out of these 25 participants, 7 participants reported more severe attacks during the moxisylyte period and 18 reported more severe attacks during the placebo period. More participants had more severe attacks during placebo treatment (P < 0.01), as reported by study authors. The remaining participants were unchanged or had less severe attacks.
Duration of attacks
Only Jaffe 1980 reported on duration of attacks. Fifteen participants recorded shorter total duration of attacks while on moxisylyte, and nine had shorter duration of attacks while on placebo (P > 0.01), as reported by study authors.
Adverse events
Le Quentrec 1991 reported no withdrawals from the trial. No evidence of a difference in adverse events was noted between the buflomedil and placebo groups (RR 1.41, 95% CI 0.27 to 7.28; 31 participants, 1 study; moderate‐certainty evidence; see Analysis 2.3). Two side effects were reported in the placebo group (gastric upset) and three in the buflomedil group (gastric burning, vertigo, hot flush). Study authors reported that side effects disappeared spontaneously with neither modification nor withdrawal of treatment.
In Jaffe 1980, one participant was withdrawn from the treatment phase because of an embolus deemed by trial authors not to be drug related, and three participants were withdrawn from the placebo phase because of side effects. A total of 13 participants during the moxisylyte phase and three during the placebo phase reported adverse events. Adverse events may be more frequent in the moxisylyte group (Analysis 2.3). Dyspepsia, heartburn, flushing, and changes in taste were reported by two or more participants in the treatment group and by none in the placebo group. No differences between subgroups were detected with the subgroup test (P = 0.28), as reported by study authors.
Neither study reported on the remaining outcomes of interest (quality of life scores, Raynaud Condition Score, or capillaroscopic flow/skin perfusion).
Prostaglandin/prostacyclin analogues
(Beraprost [ATC code B01AC19]; platelet aggregation inhibitors excluding heparin; prostacyclin analogue)
One study investigated the use of prostacyclin analogues (Vayssairat 1996). This study included 125 participants with primary RP in a parallel design and compared beraprost 40 μg three times daily with placebo. We used the phase 3 results as reported by trial authors. See summary of findings Table 3.
Frequency of attacks
Vayssairat 1996 found that the reduction of attacks from baseline was 44% in the placebo group and 37% in the beraprost group. There was no clear difference in the number of attacks per week in the beraprost group compared to the placebo group compared to the baseline values of 11 to 12 attacks per week (MD 2.0, 95% CI ‐0.35 to 4.35; 118 participants, 1 study; low‐certainty evidence; see Analysis 3.1).
Severity of attacks
Neither severity of attacks (measured on a 1 to 4‐point scale) nor disability score (measured on a 100‐mm visual analogue scale) was found to be clearly different between the beraprost group and the placebo group (MD ‐0.06, 95% CI ‐0.34 to 0.22; 118 participants, 1 study; low‐certainty evidence; see Analysis 3.2; and MD 3.00, 95% CI ‐7.48 to 13.48; 118 participants, 1 study; see Analysis 3.3, respectively). Both results had a wide confidence interval, reflecting poor precision.
Adverse events
In total, 16 participants withdrew from the study ‐ nine from the beraprost group and seven from the placebo group. Of these withdrawals, four from the beraprost group and three from the placebo group were reported to be due to side effects. Overall, more side effects were noted in the beraprost group (34/63) compared to the placebo group (21/62) (RR 1.59, 95% CI 1.05 to 2.42; 125 participants, 1 study; low‐certainty evidence; see Analysis 3.4). Headache was reported by 16 of 63 in the treatment group and by 1 of 62 in the placebo group. Two participants in each group were noted to have elevation in liver transaminases.
This study did not report on the remaining outcomes of interest (duration of attacks, quality of life scores, Raynaud Condition Score, or capillaroscopic flow/skin perfusion).
Thromboxane synthase inhibitors
(Dazoxiben, thromboxane synthase inhibitor [no ATC code])
One included study used the thromboxane synthase inhibitor dazoxiben in a cross‐over design involving 25 participants, six of whom had primary RP (Ettinger 1984). This trial compared dazoxiben 100 mg four times daily with nifedipine or placebo. Data from the nifedipine arm were not included. One participant with primary RP was withdrawn due to poor compliance. See summary of findings Table 4.
Frequency of attacks
Results show no evidence of an effect of dazoxiben compared with placebo. The mean frequency of attacks per week during intervention was 12.0 ± 6.02 (SD) in the dazoxiben group and 11.2 ± 8.11 (SD) in the placebo group (MD 0.80, 95% CI ‐1.81 to 3.41; 6 participants, 1 study; very low‐certainty evidence; see Analysis 4.1).
Adverse events
Adverse events were not reported in the subgroup analysis of participants with primary RP. This study did not report on the remaining outcomes of interest (duration of attacks, severity of attacks; quality of life scores, Raynaud Condition Score, or capillaroscopic flow/skin perfusion).
Selective serotonin reuptake inhibitors
(Ketanserin [ATC code C02KD01], other antihypertensives, serotonin antagonists)
One included study used ketanserin (Van de Wal 1987). This was a cross‐over trial comparing ketanserin 40 mg twice daily with placebo. In all, 41 participants with primary RP were included, and each treatment period duration was four weeks. See summary of findings Table 5.
Frequency of attacks
Baseline average frequency of attacks was reported to be 12 times a day. The average frequency of attacks observed with ketanserin treatment was 56 times per week versus 70 times per week in the placebo group, with SE of 7. There may be a slight reduction in the number of attacks per week in the ketanserin group compared to the placebo group (MD ‐14.0, 95% CI ‐27.72 to ‐0.28; 41 participants, 1 study; very low‐certainty evidence; see Analysis 5.1).
Severity score and subjective assessment
Severity scores were calculated by multiplying the average frequency per day by the average duration of attacks. The severity score may be reduced after ketanserin treatment compared to placebo (MD ‐133.00, 95% CI ‐162.40 to ‐103.60; 41 participants, 1 study; very low‐certainty evidence; see Analysis 5.2). A subjective assessment was also performed, with 24 participants reporting improvement in symptoms after ketanserin treatment and 14 after placebo. Later in the treatment program, 29 participants reported general improvement, and 22 of these individuals reported further improvement at the end of long‐term treatment with ketanserin.
Duration of attacks
There was no clear difference in the duration of attacks following ketanserin or placebo use (MD ‐4.00, 95% CI ‐14.82 to 6.82; 41 participants, 1 study; very low‐certainty evidence; see Analysis 5.3). Symptoms lasted approximately 50 minutes. Following treatment, the duration of these attacks lasted on average 28 minutes with both ketanserin and placebo, and both groups had an SE of six.
Capillaroscopic flow/skin perfusion
There was no clear difference between groups in digital blood flow of the second finger with treatment of 3 mmHg (MD 3.00, 95% CI ‐3.93 to 9.93; 41 participants, 1 study; very low‐certainty evidence; see Analysis 5.4).
Adverse events
Two participants withdrew from the study: one for causes unspecified, the other for non‐medical reasons. Twenty participants in the ketanserin group and 13 in the placebo group reported adverse events (RR 1.54, 95% CI 0.89 to 2.66; 41 participants, 1 study; very low‐certainty evidence; see Analysis 5.5). Headache, dry mouth, and dizziness were reported more frequently in the treatment group. This study did not report on the remaining outcomes of interest (QoL score and Raynaud Condition Score).
Nitrate/nitrate derivatives
(Organic nitrates [ATC code C01DA02])
Four studies used topical treatments of nitroglycerin or glyceryl trinitrate (GTN) (Chung 2009; Nahir 1986; Sovijarvi 1984; Teh 1995). Chung 2009 conducted a two‐week single‐blind run‐in followed by a four‐week double‐blind treatment phase. This study was a randomized, parallel, vehicle‐controlled trial. The active agent was topical 0.9% nitroglycerin in the novel vehicle called MQX‐503. Participants with primary RP included 32 of 111 in the treatment group and 37 of 108 in the placebo group. Participants were given electronic diaries to record frequency, duration, and severity of attacks and adverse experiences. They completed health assessment questionnaires (HAQs) and physicians completed global assessment questionnaires (GAQs) at each visit. Nahir 1986 studied 18 participants, seven with primary RP, with a double‐blind cross‐over design comparing Nitroderm TTS 5 mg against placebo. Participants kept a daily record of frequency and severity of attacks. Results were reported dichotomously as positive or negative responses. Sovijarvi 1984 conducted a trial on eight participants using 12.5 mg nitroglycerin versus placebo in a double‐blind cross‐over design. This study had a one‐week single‐blind run‐in period and continued as a randomized double‐blind cross‐over for two weeks. Participants kept daily diary records of frequency, severity, and duration of attacks. During the four visits of the trial, digital blood pressure responses to different temperatures were recorded. Teh 1995 recruited 42 participants including 21 with primary RP. Seventeen participants completed the trial; according to study authors, two individuals were excluded due to incomplete data of subjective assessment and two withdrew during the GTN treatment period due to side effects such as headache and nausea. After a one‐week washout period, participants were randomized to GTN 0.2 mg/hr or placebo patches for seven days. Outcomes recorded were frequency and severity of attacks. VAS was recorded at baseline and at the end of treatment. Severity of attacks was numerically scored from 0 to 3 (nil to severe). All participants had thermography recording post patches at room temperature and post cold challenge.
It was not possible to combine all data, as not all studies reported on the same outcomes, and when they did, types of data were limited. None of the studies reported quality of life scores for participants with primary Raynaud's phenomenon. See summary of findings Table 6.
Frequency of attacks
We were able to combine two studies in a meta‐analysis of frequency of attacks (Chung 2009; Teh 1995). We noted no clear overall difference in frequency between treatment and placebo groups (MD ‐1.57, 95% CI ‐4.31 to 1.17; 86 participants, 2 studies; very low‐certainty evidence; see Analysis 6.1).
In Nahir 1986, six participants reported response to Nitroderm compared to one in the placebo group.
Sovijarvi 1984 authors reported lack of differences between GTN and placebo treatments in the frequency of RP attacks without providing data.
Severity of attacks
In Teh 1995, overall severity of attacks (numbness, pain, and colour change) was reduced in the treatment group (MD ‐4.25, 95% CI ‐5.71 to ‐2.79; 17 participants, 1 study; very low‐certainty evidence; see Analysis 6.2), but the clinical relevance of this result is not clear, as there is no severity correlation for these numbers. VAS scoring was no different between groups.
In Nahir 1986, decreased severity on VAS by 2 cm was reported as a positive. Five participants reported a positive result compared to one in the placebo group.
Chung 2009 reported both participant‐ and physician‐reported that VAS scores at the target week were similar to baseline but did not report the recorded data within the article.
Sovijarvi 1984 did not report on severity of attacks.
Duration of attacks
Both Chung 2009 and Sovijarvi 1984 reported this outcome. Sovijarvi 1984 authors reported lack of differences between GTN and placebo treatments in duration of RP symptoms without providing data. Chung 2009 reported no significant difference (P > 0.2) in the duration of attacks at baseline (28.8 ± 12.9 minutes, intervention; 29.8 ± 13.6 minutes, placebo), with no significant decrease in the duration of attacks.
Capillaroscopic flow/skin perfusion
In Sovijarvi 1984, digital blood pressure analysis did not reveal any clear effect compared to placebo (MD 15.60, 95% CI ‐2.54 to 33.74; 8 participants, 1 study; see Analysis 6.3).
Teh 1995 reported no statistical differences in thermography results; however, no numerical data were provided by study authors to corroborate this statement.
Adverse events
In Chung 2009 and Nahir 1986, the numbers of adverse events were not reported separately for participants with primary Raynaud's phenomenon.
In Sovijarvi 1984, all eight participants reported headaches with nitroglycerin; four graded symptoms as severe.
In Teh 1995, two out of 21 participants with primary RP withdrew due to adverse events.
Raynaud Condition Score
Whilst Chung 2009 included participants with both primary and secondary RP, a subgroup analysis of the Raynaud Condition Score among those with primary RP was performed by study authors. This did not show clear evidence of a difference following GTN treatment (MD ‐0.36, 95% CI ‐0.98 to 0.26; 69 participants, 1 study; low‐certainty evidence; see Analysis 6.4).
None of the studies reported on quality of life.
Phosphodiesterase inhibitors
(Cilostazol [ATC Code B01AC23], platelet aggregation inhibitors excluding heparin; vardenafil [ATC Code G04BE09], drugs used in erectile dysfunction; PF‐00489791, used only in trials).
Three studies assessed the efficacy of phosphodiesterase inhibitors for improving symptoms of individuals with RP (Caglayan 2012; NCT01090492; Rajagopalan 2003). These compared vardenafil (Caglayan 2012), cilostazol (Rajagopalan 2003), and PF‐00489791 (NCT01090492), with an equivalent placebo. Two studies were cross‐over trials (Caglayan 2012; NCT01090492), and one used a parallel design (Rajagopalan 2003). In all studies, there was a mixture of primary and secondary RP. Caglayan 2012 was a two‐period cross‐over study conducted for six weeks to assess 10 mg vardenafil twice daily versus placebo. Treatments were switched following a one‐week washout phase and a follow‐up phase up to four weeks after last drug intake. Of 53 participants included in the study, six had primary RP. In Rajagopalan 2003, individuals were randomized to cilostazol 100 mg twice daily for six weeks or placebo in a double‐blind format. Forty participants completed the study ‐ 19 with primary RP. Initially, 45 individuals were screened, and two participants withdrew due to pregnancy and severe respiratory disease. Three participants who were randomly assigned to cilostazol withdrew before completion due to severe headaches. NCT01090492 measured the effects of the phosphodiesterase inhibitor, PF‐00489791 versus placebo. This study included 243 participants with primary or secondary RP. A total of 113 participants fulfilled the criteria for primary RP. Either 4 mg or 20 mg of PF‐00489791 was provided once daily for the first four‐week cross‐over period followed by placebo once daily for the second four‐week cross‐over period. A washout period of two weeks occurred between intervention periods, during which two placebo tablets matched to PF‐00489791 were given orally once daily. We report outcomes for the 20‐mg dose. See summary of findings Table 7.
Frequency of attacks
Two studies reported on this outcome, with different effects. We combined the data in a meta‐analysis (NCT01090492; Rajagopalan 2003). Overall, there was no clear reduction in the number of attacks per week (standardized mean difference [SMD] ‐0.05, 95% CI ‐6.71 to 6.61; 111 participants, 2 studies; low‐certainty evidence; see Analysis 7.1). A random‐effects analysis was used due to the different measures utilized in these studies and due to the high degree of heterogeneity (95%).
In Rajagopalan 2003, compared to placebo, cilostazol showed an increase in the frequency of attacks per week (MD 3.50, 95% CI 0.64 to 6.36; 38 participants; see Analysis 7.1). In NCT01090492, 20 mg PF‐00489791 resulted in a reduction in the frequency of Raynaud attacks compared to placebo (MD ‐3.30, 95% CI ‐3.92 to ‐2.68; 73 participants; see Analysis 7.1).
Severity of attacks
Two studies reported on this outcome and showed inconsistent results (NCT01090492; Rajagopalan 2003). We combined the data in a meta‐analysis (NCT01090492; Rajagopalan 2003). There was no clear reduction in the severity of attacks per week (SMD ‐0.03, 95% CI ‐1.04 to 0.97; 111 participants, 2 studies; very low‐certainty evidence; see Analysis 7.2). A difference was detected by the subgroup test for differences (P < 0.00001). A random‐effects analysis was used due to the different measures utilized in these studies and due to the high degree of heterogeneity (96%).
In Rajagopalan 2003, compared to placebo, cilostazol was found to cause an increase in attack severity, measured on a 1 to 9‐point scale per attack (SMD 0.50, 95% CI 0.09 to 0.91). In NCT01090492, severity scores as measured on an 11‐point Likert scale were reduced in the treatment group (SMD ‐0.53, 95% CI ‐0.63 to ‐0.43).
Duration of attacks
Only NCT01090492 reported on duration of attacks and detected no clear differences (MD ‐1.60, 95% CI ‐7.51 to 4.31; 73 participants, 1 study; low‐certainty evidence; see Analysis 7.3).
Capillaroscopic flow/skin perfusion
In Rajagopalan 2003, no clear changes in nitroglycerin‐mediated dilation or microvascular flow indexes were detected between the two groups (MD 18.00, 95% CI ‐7.34 to 43.34; 38 participants, 1 study; low‐certainty evidence; see Analysis 7.4).
Adverse events
In Caglayan 2012, the most common side effects included flushing, headache, dyspepsia, and dizziness among others. These were reported to be not specific to either group. Serious adverse events occurred in one individual receiving vardenafil; this was not related to the drug and resolved.
In Rajagopalan 2003, study authors reported that among those who completed the study, 35% of those on cilostazol complained of headaches; this was not reported in the placebo group. Two participants in the treatment group also complained of palpitations. Among both treatment and control groups, 10% reported gastrointestinal side effects such as diarrhea and nausea.
NCT01090492 reported that 34 out of 54 participants in the 20‐mg treatment group and 43 out of 102 participants in the placebo group experienced adverse events, resulting in an RR of 1.49 (Analysis 7.5). Headache was the most commonly reported side effect, affecting 14 participants in the 20‐mg treatment group and 9 in the placebo group.
Raynaud Condition Score
Two studies reported on RCS (Caglayan 2012; NCT01090492). A meta‐analysis revealed no clear reduction (SMD ‐0.80, 95% CI ‐1.74 to 0.13; 79 participants, 2 studies; low‐certainty evidence; see Analysis 7.6). A random‐effects analysis was used due to the different measures used in these studies and due to the high degree of heterogeneity (52%). Testing for subgroup differences was performed (P = 0.15).
Caglayan 2012 reported subgroup data for primary RP. Among six individuals with primary RP, a small reduction in RCS was noted (MD ‐1.61, 95% CI ‐3.09 to ‐0.13; 6 participants). Study authors reported a long‐term benefit of vardenafil, evidenced by lower RCS during the washout phase and in the second phase of the study among individuals who were initially assigned to vardenafil.
In NCT01090492, 20 mg PF‐00489791 shows improved RCS (MD ‐0.52; 95% CI ‐0.60 to ‐0.44; 73 participants).
None of the studies reported on quality of life.
Discussion
Summary of main results
This review summarizes the latest evidence for treatment of primary Raynaud's phenomenon (RP) using drugs with vasodilator effects excluding calcium channel blockers. Seven new studies were identified for this update. In total, 15 studies with 635 participants were included. The studies involved ten drugs from eight drug classes: angiotensin‐converting enzyme (ACE) inhibitors, 5‐HT2 receptor antagonists, nitrates/nitrate derivatives, phosphodiesterase inhibitors, prostacyclin analogues, alpha blockers, and thromboxane synthase inhibitors ‐ all compared to placebo. Of the seven new studies, four investigated the use of glyceryl trinitrate (GTN) and three investigated the use of phosphodiesterase inhibitors. This update also aimed to include intravenous and topical formulations.
Angiotensin‐converting enzyme (ACE) inhibitors
Three studies investigated the use of ACE inhibitors (Challenor 1991; Madsen 1984; Rustin 1987); these studies demonstrated that there may be a small increase in the frequency of attacks with treatment. They provided no evidence of a difference in severity of attacks, duration of attacks, or adverse events. Quality of life (QoL) and Raynaud Condition Score (RCS) were not reported. As Challenor 1991 was heavily weighted (95.6%) in the analysis, the overall result was similar to the result in Challenor 1991, despite heterogeneous outcomes in Madsen 1984 and Rustin 1987. See summary of findings Table 1.
Captopril alone did not improve frequency of attacks, severity of attacks, or duration of attacks, nor did it improve digital blood flow (Madsen 1984; Rustin 1987).
Enalapril alone is associated with a small increase in the frequency of attacks per week and worsening of subjective assessment of improvement on a 10‐point scale (no clear difference in severity rating of attacks was noted) (Challenor 1991).
Alpha blockers
Two studies used alpha blockers. One study investigated buflomedil (Le Quentrec 1991), and one moxisylyte (Jaffe 1980). We were not able to combine the data in a meta‐analysis because of how the results were presented. Buflomedil probably reduced the frequency of attacks per week compared to placebo. This effect was small but precise. Results showed a small improvement in severity scores between buflomedil and placebo (Le Quentrec 1991).
An effect in favor of moxisylyte compared with placebo was reported by study authors, along with numbers of participants experiencing fewer attacks (P < 0.02), less severe symptoms (P < 0.01), and attacks of shorter duration (). The study comprised a small number of participants, and the magnitude of the effect was not clear.
No evidence shows a difference in adverse events between buflomedil and placebo groups. More side effects were observed among participants in the moxisylyte group compared to the placebo group. See summary of findings Table 2.
Prostaglandin/prostacyclin analogues
One study investigated the use of prostacyclin analogues, comparing beraprost with placebo (Vayssairat 1996). Results show no evidence of benefit of beraprost compared with placebo for frequency of attacks, severity of attacks, nor disability score (Vayssairat 1996). Overall, more side effects were noted in the beraprost group (34/64) compared to the placebo group (21/62). This study did not report on duration of attacks, quality of life (QoL), or Raynaud Condition Score (RCS). See summary of findings Table 3.
Thromboxane synthase inhibitors
One included study compared the thromboxane synthase inhibitor dazoxiben versus placebo (Ettinger 1984). Results show no evidence of beneficial effects of dazoxiben compared with placebo for frequency of attacks. Adverse events were not reported in the subgroup analysis of participants with primary RP, and the study did not report on duration of attacks, severity of attacks, QoL, or RCS. See summary of findings Table 4.
Selective serotonin reuptake inhibitors
One included study compared ketanserin with placebo (Van de Wal 1987). There may be a slight reduction in the number of attacks per week with ketanserin compared to placebo. Similarly, the severity score may be slightly reduced after ketanserin treatment. No evidence suggests that ketanserin reduced the duration of attacks, or that adverse events were increased in either group. This study did not report on QoL or on RCS. See summary of findings Table 5.
Nitrate/nitrate derivatives
Four studies compared topical treatment of nitroglycerin or glyceryl trinitrate (GTN) versus placebo (Chung 2009; Nahir 1986; Sovijarvi 1984; Teh 1995). The meta‐analysis for topical GTN showed no evidence of reduced frequency of attacks per week in two highly heterogeneous studies. Nahir 1986 utilized dichotomous outcomes to demonstrate an effect on reduction in frequency of attacks with topical GTN. However, the clinical implication was difficult to judge due to the small sample size of the study and lack of qualitative data. Sovijarvi 1984 reported lack of differences between GTN and placebo in the frequency of RP attacks without providing data. Three studies reported on severity of attacks. In Teh 1995, overall severity of attacks (numbness, pain, and color change) was reduced in the treatment group, but the clinical relevance of this result is not clear. In Nahir 1986, five participants from the GTN group reported a positive result compared to one in the placebo group. Chung 2009 recorded severity of attacks reported by both physicians and patients on a 0 to 100 scale; no significant difference was noted in either group, and no numeric data were reported. Sovijarvi 1984 reported on duration of attacks but observed lack of differences between GTN and placebo (no data provided). Chung 2009 reported no significant differences between groups in the duration of attacks. Chung 2009 reported RCS but did not show evidence of a difference. None of the studies reported on QoL. See summary of findings Table 6.
Phosphodiesterase inhibitors
Three studies assessed the efficacy of phosphodiesterase inhibitors for improving symptoms among individuals with RP (Caglayan 2012; NCT01090492; Rajagopalan 2003). These studies compared vardenafil (Caglayan 2012), cilostazol (Rajagopalan 2003), and PF‐00489791 (NCT01090492), with an equivalent placebo. Meta‐analysis showed no evidence of a difference between treatment and placebo in terms of frequency of attacks or severity of attacks (NCT01090492; Rajagopalan 2003). Only NCT01090492 reported on duration of attacks; no clear difference was detected. Rajagopalan 2003 authors reported that 35% of participants on cilostazol complained of headaches, and this was not reported in the placebo group. In NCT01090492, 34 of 54 participants in the treatment arm experienced adverse events, as did 43 of 102 participants in the placebo arm. Headache was the most common adverse event among 14 participants in the treatment arm and among 9 participants in the placebo arm. There was no clear reduction in RCS (Caglayan 2012; NCT01090492). None of the studies reported on QoL See summary of findings Table 7.
The following classes of drugs, or single drugs with vasodilator effects, were not represented in the included studies: neprilysin inhibitors, angiotensin‐II receptor antagonists, potassium channel activators, reflex sympathetic stimulators, endothelin antagonists, naftidrofuryl, and antihistamines.
Overall completeness and applicability of evidence
It is difficult to quantify the clinical effects of various treatments. Chung 2009 and Vayssairat 1996 performed sample size calculations; Chung 2009 aimed to find at least a 20% difference in the proportion of participants with 30% improvement in RCS. Vayssairat 1996 was powered to detect a 50% difference in the number of attacks, irrespective of placebo effect, which would be a clinically meaningful outcome for patients with RP. Study authors acknowledged the absence of a defined minimally detectable clinical difference for RCS.
As mentioned in the Included studies section, significant changes have been made to the definition of primary RP over the last 30 years; this, in part, may affect the observed heterogeneity of outcomes. As consensus on criteria was reached as of 2014, this is less likely to be an issue with new studies. Variation in demographics such as gender and smoking and disease severity at baseline may also play a role in heterogeneity of outcomes. The consistency of winter months selected for the study season was reasonable, but as temperature may differ in different parts of the world or even among study settings, reporting of average temperatures within settings may aid in standardization of investigation of primary RP and may further reduce heterogeneity and bias. Of note, only one of the included studies provided a breakdown of the demographics of study participants; educational level, occupation, and other socio‐demographic factors were not discussed in all studies.
Early results are promising, especially in large cohort, multicenter trials (NCT01090492), for certain agents such as topical glyceryl trinitrate and phosphodiesterase inhibitors. The review highlights various agents that have been studied, but there is a paucity of data to suggest a clinically meaningful outcome with these therapies.
Although Nahir 1986 included a female population only, we believe that the broad range of patient demographics in other studies allows for general applicability, thus providing external validity. We acknowledge additional difficulties in that RP diagnostic criteria have changed several times over the years. A predominantly outpatient setting during winter was appropriate for studying RP with these study designs.
Although it was the intention of review authors to incorporate the degree of sonographic or capillaroscopic findings as a demonstration of the effects of therapy, too few trials were included for performance of a meta‐analysis.
As mentioned in previous reviews, there were difficulties in interpretation and analysis due to lack of simplistic mean and standard deviation and/or standard error format, or lack of reported data, which rendered it difficult to calculate outcomes for meta‐analysis despite review authors' intentions.
Quality of the evidence
Some reviews reported outcome measures associated with significant P values, indicating that the effect seen was unlikely due to chance. Despite P values, the magnitude of effects was generally small. In addition, imprecision along with evidence of very low to moderate certainty makes it difficult to draw definitive conclusions regarding treatment outcomes.
Although meta‐analyses for phosphodiesterase inhibitors, ACE inhibitors, and glyceryl trinitrate were performed, we acknowledge that the outcomes were non‐specific in nature, and further study of individual drugs in this class is needed, as evidenced by differences in outcomes (e.g. cilostazol versus PF‐00489791). We also note that there is inherent heterogeneity when meta‐analyses of drug classes are performed, as the pharmacokinetics and the pharmacodynamics of an agent can be significantly different and may be impacted by study design, leading to differences in outcomes. Additionally, topical drug delivery shows heterogeneity, as delivery vehicle, patch design, concentration of agent, and duration and method of application would impact differences in effects between studies. Most studies ‐ 11 out of 15 ‐ used a cross‐over design, which provides some advantages for studies with small sample sizes and a parallel design. As participants are involved in both control and treatment groups, a larger pool of data can be analyzed, increasing the power of study outcomes. Results are more precise, as interventions are evaluated among the same participants, which reduces the effects of demographic factors. In this review, however, data had to be presented as though studies used a parallel design due to inadequately reported results. This subsequently led to loss of power and precision from what was intended as originally designed. Some studies did not have any washout periods between carry‐over periods (Jaffe 1980; Madsen 1984; Rustin 1987; Sovijarvi 1984; Teh 1995). However, elimination times were less than the period between last treatment and first dose given after carry‐over. Remaining studies with a cross‐over design had adequate washout periods, which exceeded elimination times of treatment (Caglayan 2012; Challenor 1991; Ettinger 1984; Nahir 1986; NCT01090492; Van de Wal 1987).
Jaffe 1980 and Nahir 1986 reported dichotomous outcomes that were difficult to clinically quantify given the lack of standardization. Examples of these include "patients experienced shorter attacks" and "patients experienced more severe attacks." Therefore, clinical importance was not assessed.
All studies reported subjective outcomes of Raynaud's phenomenon, which were the main defined outcomes. Primary RP rarely causes permanent tissue damage. Hence, the aim of treatment has been prophylactic and symptomatic control. Only three studies ‐ Caglayan 2012; Chung 2009; NCT01090492 ‐ reported a validated outcome measure ‐ the Raynaud Condition Score, which has been used as the primary outcome measure for clinical trials for RP since the early 2000s (Merkel 2002). Apart from this, there was considerable variation in types of severity measurements that did not include the same scales. Variability in reporting of subjective measures made comparison of results among studies difficult or impossible. One of the secondary outcomes in this review was capillaroscopic flow/skin perfusion by ultrasound scan, which was reported in four studies (Rajagopalan 2003; Rustin 1987; Sovijarvi 1984; Van de Wal 1987). This may be a surrogate measure representing the severity of disease activity. However, the clinical correlation is unclear, as variability in measurements made it impossible to compare data and assess clinical importance.
The certainty of evidence for outcomes lies between very low and low for the reasons described above, except for outcomes reported in Le Quentrec 1991, for which evidence was of moderate certainty. A number of outcomes relied on data from one or two studies, and so evidence was downgraded due to small numbers of participants, imprecision, and bias concerns. Heterogeneity among methodological reasons and inconsistency in the direction of effect also led to downgrading of the evidence.
Although overall sample size was increased from that in previous version of the review, lack of comparable outcomes and bias concerning imprecision and heterogeneity among the same classes of medications resulted in low‐certainty evidence. Overall, It is not possible to assess the clinical implications of effects of various treatments for this reason. There were suggestions that treatment may provide some benefit for subjective measures of RP; however, the magnitude of effect observed was small, and thus clinical relevance is not clear.
Potential biases in the review process
A systematic approach to the search for studies for inclusion in this Cochrane Review was intended to reduce bias; this included defining inclusion and exclusion criteria and limiting study design to randomized controlled trials (RCTs). Including only studies that compared vasodilating drugs with placebo enhanced efficacy as reported in the results and provided a baseline for all treatment comparisons. Furthermore, the randomization aspect generally favors lower risk of bias compared to other study designs and minimizes confounding factors. However, when only RCTs are included, generalizing results to the general population becomes more difficult for a condition that can be affected by many individual and environmental factors; therefore the data may not represent individual outcomes.
Within the method itself, both cross‐over and parallel studies were included. Each design has its advantages, with cross‐over allowing evaluation of interventions in the same group of participants, reducing long‐term participant response bias. Parallel studies generally utilize a larger, and consequently more diverse, participant pool. However, it is noted that due to different aspects of study design used by respective researchers, such as washout time frames and carry‐over, bias in the review process may have been introduced when these studies were compared against one another, and when data were presented in a way that allows best comparison between parallel and cross‐over studies.
Many studies were described as using "double‐blinding" through utilization of identical placebo without elaboration of the manner in which blinding was applied. This contravenes the CONSORT 2010 statement regarding double blinding terminology and the recommendation for explicit distinction of blinded parties (i.e. assessors, participants, or investigators). We therefore deemed these studies to be at unclear risk of bias.
For the purposes of this review, only data from participants with primary RP were included. Combined data from participants with primary and secondary RP were not included in this review. Although this does discriminate findings against the broader group of individuals who may also have secondary RP, the results are more accurate for the group of individuals who experienced attacks according to current diagnostic criteria. Consequently, there were fewer independent rheumatologic pathologic variables that could have impacted participant response to treatment than would likely occur in secondary RP, by definition. Similarly, through inclusion of only vasodilatory medications without calcium channel blocker properties and by exclusion of alternative medicines and non‐pharmacologic modalities, a more thorough analysis of efficacy among treatment groups could be performed. However, it is appreciated that again, the results are not inclusive of or applicable to patients on multiple therapies or receiving different therapies in the community setting. Some studies include both primary and secondary RP.
Through inclusion only of trials that lasted longer than one week, bias masked by duration was controlled. The specific time frame of trials allowed for processing of more longitudinal data and enabled assessment of whether therapy was effective in altering outcome measures over the long term ‐ a feature more classically applied to clinical methods of treatment.
Despite inclusion of alternative routes of administration, thus increasing the sample size from the previous review of 290 participants in eight included studies to 635 participants in 15 studies, the final number of included studies was comparably limited. The review suffered from two primary limitations: first, the multitude of single studies on different drugs, and second, the high variability in study design and outcome measures. It may be argued that through use of specific inclusion and exclusion criteria, there was selection bias with potential for outcome bias during analysis of results, and that the quality of other excluded studies may have been methodologically superior. There is no way to ensure that certain studies were not missed with the search terminologies used. Although the approach taken was to narrow down studies, it is appreciated that certain studies that were ultimately testing the same outcomes in the included patient population but were published using different vocabulary may have been missed. By not restricting published language or study origin, cultural bias was reduced, and this increased external validity for multicultural general populations. As discussed above, however, allowing inclusion of specific studies that treated only one population sample, for example, women, may have introduced risk of overall application of results for the general community.
Most outcome measures assessed were subjective in nature. Given that primary RP, outside of its attacks, is not commonly known to present with pathologic features or damage, and that attacks are experienced by affected patients, the measures presented are appropriate in measuring any response to treatment. Possible risk of bias from subjective (i.e. self‐reported) outcomes has been taken into consideration. More recent studies use more objective measures, such as capillaroscopic flow/skin perfusion measurements, which were not used in older studies; it is only in recent times that level of blood flow or amount of vasoconstriction is postulated to directly correlate with symptom and subjective attack experience (Wilkinson 2018). This reduces the comparability of findings.
Not limiting year of publication allows for consideration of an increased number of studies fitting the inclusion criteria. However, this introduces several inherent risks. First, in current practice, during application for approval to proceed with trials and studies, a strict and widely used code of ethics and principles must be approved before a study can occur. As 32 years has passed between oldest and newest publications ‐ Jaffe 1980 and Caglayan 2012 ‐ it is probable that quality of methods and ethical limitations may vary between trials of the 1980s and studies of the 2010s. Although this cannot be proved or quantified due to lack of available trial information and details regarding approval methods, it suggests that there is risk of bias with regard to external validity.
Agreements and disagreements with other studies or reviews
The results of this review do not provide evidence in favor of vasodilator use ‐ oral or topical ‐ for treatment of primary RP. This is consistent with reports from other reviews on treatment of primary RP (Distler 2006; Pope 2011).
Distler 2006 included four studies looking at effects of vasodilators apart from calcium channel blockers (CCBs) on primary RP, which were included in our review (Caglayan 2012; Challenor 1991; Rustin 1987; Vayssairat 1996). The Distler review included studies on secondary RP, as well as use of CCBs. Defined outcomes were different, and changes in fingertip ulcers and peripheral circulation were described. Pope 2011 included a total of 20 studies on effects of drugs for primary RP. Fifteen studies compared CCBs versus placebo, and two studies looked at the effect of Inositol nicotinate, which is not a vasodilator. Among the remaining three studies, two were excluded from our review as it is not possible to distinguish participants with primary and secondary RP (Davinroy 1993; Wollersheim 1986b). Data on moxisylyte versus placebo were included in Pope 2011, for which the author agrees that evidence is of very low quality based on the GRADE system.
Our review included RCTs looking at effects of different vasodilators on primary RP excluding CCBs. We excluded a large number of studies because it is not possible to differentiate participants with primary and secondary RP. We excluded participants with secondary RP because of differences in pathophysiology and clinical complications associated with primary and secondary RP. Primary RP rarely leads to permanent tissue damage, whereas secondary RP is associated with abnormalities in vascular structure that may be irreversible and can lead to digital ulceration (Herrick 2005). Defined outcomes were mainly subjective for this reason. This review performed a GRADE evaluation of the certainty of evidence for interventions.
Study flow diagram.
Risk of bias graph: review authors' judgements about each risk of bias item presented as percentages across all included studies.
Risk of bias summary: review authors' judgements about each risk of bias item for each included study.
Comparison 1: ACE inhibitor versus placebo, Outcome 1: Frequency of attacks per week
Comparison 1: ACE inhibitor versus placebo, Outcome 2: Severity of symptoms
Comparison 1: ACE inhibitor versus placebo, Outcome 3: Duration of attacks
Comparison 1: ACE inhibitor versus placebo, Outcome 4: Capillaroscopic flow/skin perfusion by ultrasound
Comparison 1: ACE inhibitor versus placebo, Outcome 5: Subjective assessment of improvement (10‐cm visual analogue scale)
Comparison 1: ACE inhibitor versus placebo, Outcome 6: Adverse events
Comparison 2: Alpha blockers versus placebo, Outcome 1: Frequency of attacks ‐ buflomedil
Comparison 2: Alpha blockers versus placebo, Outcome 2: Severity of symptoms ‐ buflomedil
Comparison 2: Alpha blockers versus placebo, Outcome 3: Adverse events ‐ buflomedil
Comparison 3: Beraprost versus placebo, Outcome 1: Frequency of attacks per week
Comparison 3: Beraprost versus placebo, Outcome 2: Severity of attacks (1 to 4 scale)
Comparison 3: Beraprost versus placebo, Outcome 3: Disability (100‐mm visual analogue scale)
Comparison 3: Beraprost versus placebo, Outcome 4: Adverse events
Comparison 4: Dazoxiben versus placebo, Outcome 1: Frequency of attacks per week
Comparison 5: Ketanserin versus placebo, Outcome 1: Frequency of attacks per week
Comparison 5: Ketanserin versus placebo, Outcome 2: Severity score (frequency of attacks/d × duration of attacks)
Comparison 5: Ketanserin versus placebo, Outcome 3: Duration of attacks per day (minutes)
Comparison 5: Ketanserin versus placebo, Outcome 4: Capillaroscopic flow/skin perfusion by ultrasound
Comparison 5: Ketanserin versus placebo, Outcome 5: Adverse events
Comparison 6: Nitrate/nitrate derivatives versus placebo, Outcome 1: Frequency of attacks
Comparison 6: Nitrate/nitrate derivatives versus placebo, Outcome 2: Severity of symptoms
Comparison 6: Nitrate/nitrate derivatives versus placebo, Outcome 3: Capillaroscopic flow/skin perfusion by ultrasound
Comparison 6: Nitrate/nitrate derivatives versus placebo, Outcome 4: Raynaud Condition Score
Comparison 7: Phosphodiesterase inhibitors versus placebo, Outcome 1: Frequency of attacks per week
Comparison 7: Phosphodiesterase inhibitors versus placebo, Outcome 2: Severity of symptoms
Comparison 7: Phosphodiesterase inhibitors versus placebo, Outcome 3: Duration of attacks
Comparison 7: Phosphodiesterase inhibitors versus placebo, Outcome 4: Capillaroscopic flow/skin perfusion by ultrasound
Comparison 7: Phosphodiesterase inhibitors versus placebo, Outcome 5: Adverse events
Comparison 7: Phosphodiesterase inhibitors versus placebo, Outcome 6: Raynaud Condition Score
| ACE inhibitors compared to placebo for treatment of primary Raynaud's phenomenon | ||||||
| Patient or population: primary Raynaud's phenomenon | ||||||
| Outcomes | Anticipated absolute effects* (95% CI) | Relative effect | №. of participants | Certainty of the evidence | Comments | |
|---|---|---|---|---|---|---|
| Risk with placebo | Risk with ACE inhibitor | |||||
| Frequency of attacks (per week) (4 to 6‐week follow‐up) | MD 0.79 higher | ‐ | 44 (3 RCTs) | ⊕⊕⊝⊝ | There may be increased frequency of attacks with treatment | |
| Severity of attacks (mild, moderate, and severe. then converted to numeric representation) (4 to 6‐week follow‐up) | MD 0.17 lower | ‐ | 34 (2 RCTs) | ⊕⊕⊝⊝ | No evidence of a difference | |
| Duration of attacks (minutes) (4 to 6‐week follow‐up) | MD 0.54 higher (1.34 lower to 2.42 higher) | ‐ | 14 (1 RCT) | ⊕⊕⊝⊝ | No evidence of a difference | |
| QoL | See comment | ‐ | ‐ | ‐ | This outcome was not reported by any study | |
| Adverse events (4 to 6‐week follow‐up) | 182 per 1000 | 245 per 1000 (122 to 496) | RR 1.35 (95% CI 0.67 to 2.73) | 46 (3 RCTs) | ⊕⊕⊝⊝ | No evidence of a difference |
| RCS | See comment | ‐ | ‐ | ‐ | This outcome was not reported by any study | |
| *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. | ||||||
| aOutpatient clinic ‐ Madsen 1984 and Challenor 1991 did not specify. | ||||||
| Alpha blockers compared to placebo for primary Raynaud's phenomenon | ||||||
| Patient or population: primary Raynaud's phenomenon | ||||||
| Outcomes | Anticipated absolute effects* (95% CI) | Relative effect | №. of participants | Certainty of the evidence | Comments | |
|---|---|---|---|---|---|---|
| Risk with placebo | Risk with alpha blocker | |||||
| Frequency of attacks (4 weeks' to 180 days' follow‐up) | Moxisylyte | ‐ | 33 (1 RCT) | ‐ | 19 participants had fewer attacks during the moxisylyte period and 10 during the placebo period. Four participants had an equal number of attacks in each period | |
| See comment | ||||||
| Buflomedil | ‐ | 31 (1 RCT) | ⊕⊕⊕⊝ | Frequency of attacks per week may be reduced with buflomedil treatment | ||
| MD 8.82 lower | ||||||
| Severity of attacks (Dichotomous outcome over 4 weeks' to 180 days' follow‐up) | Moxisylyte | ‐ | 25 (1 RCT) | ‐ | Of 25 participants, 7 reported more severe attacks during the moxisylyte period and 18 participants reported more severe attacks during the placebo period | |
| See comment | ||||||
| Buflomedil | ‐ | 31 (1 RCT) | ⊕⊕⊝⊝ | Severity of attacks may be slightly reduced | ||
| MD 0.41 lower | ||||||
| Duration of attacks (4 weeks) | Moxisylyte | ‐ | 33 | ‐ | Fifteen participants recorded shorter total duration of attacks while on moxisylyte, and 9 had shorter duration of attacks on placebo | |
| See comment | ||||||
| Buflomedil | ‐ | ‐ | ‐ | This outcome was not reported | ||
| See comment | ||||||
| QoL | See comment | ‐ | ‐ | ‐ | This outcome was not reported | |
| Adverse events (4 weeks' to 180 days' follow‐up) | Moxisylyte | ‐ | 33 | ‐ | Total of 13 participants during the moxisylyte phase and 3 during the placebo phase reported adverse events | |
| See comment | ||||||
| Buflomedil | RR 1.41 | 31 (1 RCT) | ⊕⊕⊕⊝ | No evidence of a difference in adverse events was seen between buflomedil and placebo groups | ||
| 133 per 1000 | 188 per 1000 | |||||
| RCS | See comment | ‐ | ‐ | ‐ | This outcome was not reported | |
| *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. | ||||||
| aBuflomedil (in Le Quentrec 1991) and moxisylyte (thymoxamine; in Jaffe 1980). | ||||||
| Beraprost compared to placebo for primary Raynaud's phenomenon | ||||||
| Patient or population: primary Raynaud's phenomenon | ||||||
| Outcomes | Anticipated absolute effects* (95% CI) | Relative effect | №. of participants | Certainty of the evidence | Comments | |
|---|---|---|---|---|---|---|
| Risk with placebo | Risk with beraprost | |||||
| Frequency of attacks per week (6 weeks' follow‐ up) | MD 2 higher | ‐ | 118 (1 RCT) | ⊕⊕⊝⊝ | No evidence of a difference | |
| Severity of attacks (1 to 4 scale, 6 weeks' follow‐up) | MD 0.06 lower | ‐ | 118 | ⊕⊕⊝⊝ | No evidence of a difference | |
| Duration of attacks | See comment | ‐ | ‐ | ‐ | This outcome was not reported | |
| QoL | See comment | ‐ | ‐ | ‐ | This outcome was not reported | |
| Adverse events (6 weeks' follow‐up) | Study population | RR 1.59 | 125 | ⊕⊕⊝⊝ | There may be more adverse events in the beraprost group | |
| 339 per 1000 | 539 per 1000 | |||||
| RCS | See comment | This outcome was not reported | ||||
| *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. | ||||||
| aWe downgraded by one step due to inconsistency. | ||||||
| Dazoxiben compared to placebo for primary Raynaud's phenomenon | ||||||
| Patient or population: primary Raynaud's phenomenon | ||||||
| Outcomes | Anticipated absolute effects* (95% CI) | Relative effect | №. of participants | Certainty of the evidence | Comments | |
|---|---|---|---|---|---|---|
| Risk with placebo | Risk with dazoxiben | |||||
| Frequency of attacks Per week (over 2 weeks) | MD 0.8 higher | ‐ | 6 (1 RCT) | ⊕⊝⊝⊝ | There was no evidence of an effect of dazoxiben compared with placebo | |
| Severity of attacks | See comment | ‐ | ‐ | ‐ | This outcome was not reported | |
| Duration of attacks | See comment | ‐ | ‐ | ‐ | This outcome was not reported | |
| QoL | See comment | ‐ | ‐ | ‐ | This outcome was not reported | |
| Adverse events | See comment | ‐ | ‐ | ‐ | Unclear how many participants had primary Raynaud's phenomenon among 5 participants who had adverse events | |
| RCS | See comment | ‐ | ‐ | ‐ | This outcome was not reported | |
| *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. | ||||||
| aWe downgraded by three steps due to risk of bias concerns (selection and performance bias), inconsistency of result, and imprecision (small number of participants with primary Raynaud's phenomenon). | ||||||
| Ketanserin compared to placebo for primary Raynaud's phenomenon | ||||||
| Patient or population: primary Raynaud's phenomenon | ||||||
| Outcomes | Anticipated absolute effects* (95% CI) | Relative effect | №. of participants | Certainty of the evidence | Comments | |
|---|---|---|---|---|---|---|
| Risk with placebo | Risk with ketanserin | |||||
| Frequency of attacks Per week (over 4 weeks) | MD 14 lower | ‐ | 41 (1 RCT) | ⊕⊝⊝⊝ | There may be a slight reduction in the number of attacks per week in the ketanserin group compared to the placebo group | |
| Severity of attacks (frequency of attacks/d × duration of attacks, over 4 weeks) | MD 133 lower | ‐ | 41 (1 RCT) | ⊕⊝⊝⊝ | Severity score may be slightly reduced after ketanserin compared to placebo | |
| Duration of attacks Per day (minutes, over 4 weeks) | MD 4 lower | ‐ | 41 (1 RCT) | ⊕⊝⊝⊝ | No evidence of a difference in duration of attacks | |
| QoL | See comment | ‐ | ‐ | ‐ | This outcome was not reported | |
| Adverse events (over 4 weeks) | Study population | RR 1.54 | 41 | ⊕⊝⊝⊝ | Headache, dry mouth. and dizziness were reported more frequently in the treatment group | |
| 317 per 1000 | 488 per 1000 | |||||
| RCS | See comment | ‐ | ‐ | ‐ | This outcome was not reported | |
| *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. | ||||||
| aWe downgraded by three steps due to risk of bias concerns (selection and performance bias), imprecision (small number of participants), and inconsistency (wide confidence intervals). | ||||||
| Nitroglycerin or glyceryl trinitrate (GTN) compared to placebo for primary Raynaud's phenomenon | ||||||
| Patient or population: primary Raynaud's phenomenon | ||||||
| Outcomes | Anticipated absolute effects* (95% CI) | Relative effect | №. of participants | Certainty of the evidence | Comments | |
|---|---|---|---|---|---|---|
| Risk with placebo | Risk with GTN | |||||
| Frequency of attacks (1 or 4 weeks' follow‐up) | Continuous data | ‐ | 86 (2 RCTs) | ⊕⊝⊝⊝ | No evidence of a difference | |
| MD 1.57 lower | ||||||
| Dichotomous data | ‐ | 14 | ‐ | Six participants reported response to Nitroderm compared to 1 in the placebo group | ||
| See comment | ||||||
| Severity of attacks (1 week follow‐up) | Continuous data | ‐ | 17 (1 RCT) | ⊕⊝⊝⊝ | Severity may be reduced but the clinical relevance of this is unclear | |
| MD 4.25 lower | ||||||
| Dichotomous data | ‐ | 14 | ‐ | Five participants reported a positive result compared to 1 in the placebo group | ||
| See comment | ||||||
| Duration of attacks (2 or 4 weeks' follow‐up) | See comment | ‐ | 77 (2 RCTs) | ‐ | Sovijarvi 1984 reported lack of differences in duration of attacks between GTN and placebo groups (no data provided) Chung 2009 reported no significant decrease in duration of attacks | |
| QoL | See comment | ‐ | ‐ | ‐ | No studies reported on this outcome | |
| Adverse events (3 weeks' follow‐up) | See comment | ‐ | 8 | ‐ | All 8 participants reported headaches with nitroglycerin | |
| RCS (4 weeks' follow‐up) | MD 0.36 lower | ‐ | 69 (1 RCT) | ⊕⊕⊝⊝ | No evidence of a difference following GTN treatment | |
| *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. | ||||||
| aStudies were conducted in outpatient settings, except for Sovijarvi 1984 (laboratory room). | ||||||
| Phosphodiesterase inhibitors compared to placebo for primary Raynaud's phenomenon | ||||||
| Patient or population: primary Raynaud's phenomenon | ||||||
| Outcomes | Anticipated absolute effects* (95% CI) | Relative effect | №. of participants | Certainty of the evidence | Comments | |
|---|---|---|---|---|---|---|
| Risk with placebo | Risk with phosphodiesterase inhibitors | |||||
| Frequency of attacks (per week) (4 and 6 weeks' follow‐up) | SMD 0.05 lower | ‐ | 111 (2 RCTs) | ⊕⊕⊝⊝ | No evidence of a difference but studies showed conflicting results | |
| Severity of attacks (Likert scale 1 to 9 or 11 points; 4 and 6 weeks' follow‐up) | SMD 0.03 lower | ‐ | 111 (2 RCTs) | ⊕⊝⊝⊝ | No evidence of a difference but studies showed conflicting results | |
| Duration of attacks (4 weeks' follow‐up) | MD 1.60 lower (7.51 lower to 4.31 higher) | ‐ | 73 (1 RCT) | ⊕⊕⊝⊝ | No evidence of a difference | |
| QoL | See comment | ‐ | ‐ | ‐ | This outcome was not reported | |
| Adverse events | See comment | ‐ | 138 (3 RCTs) | ‐ | 35% of participants on cilostazol complained of headaches. Two participants on cilostazol complained of palpitations These were not reported in the placebo group No specific adverse events were reported in the vardenafil group compared to the placebo group 34/54 participants experienced adverse events in the PF‐00489791 treatment group compared with 43/102 participants in the placebo group. Headache was the most commonly reported adverse event, affecting 14 participants in the PF‐00489791 group and 9 participants in the placebo group | |
| RCS (6 weeks' follow‐up) | SMD 0.80 lower | ‐ | 79 (2 RCTs) | ⊕⊕⊝⊝ | No evidence of a difference | |
| *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. | ||||||
| aStudies investigated cilostazol (Rajagopalan 2003); vardenafil (Caglayan 2012); and PF‐00489791 (NCT01090492). | ||||||
| Pharmacologic class | ATC code | Generic name |
|---|---|---|
| Phosphodiesterase‐3 inhibitor | C01CE02, B01AC23 | Milrinone, cilostazol |
| Neprilysin inhibitor | C09DX04 | Sacubitril |
| Nitrate/nitrate derivative | C01D, C08EX02, C02DD01 | Glyceryl trinitrate, isosorbide dinitrate, isosorbide mononitrate, nicorandil, sodium nitroprusside |
| Angiotensin‐converting enzyme (ACE) inhibitor | C09A | Captopril, enalapril, fosinopril, lisinopril, perindopril, quinapril, ramipril, trandolapril |
| Angiotensin‐II receptor antagonist | C09C | Candesartan, eprosartan, irbesartan, losartan, olmesartan, telmisartan, valsartan |
| Selective alpha‐blocker | C02CA, C02AC | Prazosin, terazosin, moxonidine, tamsulosin, alfuzosin |
| Potassium channel activator | C02DA | Diazoxide |
| Reflex sympathetic stimulator | C02D | Hydralazine, minoxidil |
| Prostacyclin/prostanoid | B01AC, G04BE01 | Epoprostenol, iloprost, treprostinil, alprostadil, beraprost |
| Endothelin antagonist | C02KX | Ambrisentan, bosentan, macitentan, rociguat |
| Phosphodiesterase‐5 inhibitor | G04BE | Sildenafil, tadalafil, vardenafil |
| Selective serotonin reuptake inhibitor | N06AB, C02KD01 | Citalopram, dapoxetine, escitalopram, fluoxetine, fluvoxamine, paroxetine, sertraline, ketanserin |
| Peripheral vasodilator, mechanism not well defined | C04AD | Pentoxifylline, perhexiline |
| AMH: Australian Medicines Handbook. | ||
| AMH Chapter | Class | ATC Code | Trials |
|---|---|---|---|
| Cardiovascular drugs | Vasodilators, antihypertensive drugs, and prostacyclin analogues | C02D, | Included: Rajagopalan 2003; Vayssairat 1996 Ongoing: NCT02583789 Awaiting classification: Sakaguchi 1990 |
| Cardiovascular drugs | Phosphodiesterase type 5 inhibitors | G04BE | Included: Caglayan 2012; NCT01090492 Ongoing: EUCTR2005‐000295‐41‐DE |
| Cardiovascular drugs/endocrine drugs | Alpha adrenoreceptor‐blocking drugs | C02C | Included: none |
| Unlisted | Serotonin antagonists | C02KD | Included: Van de Wal 1987 |
| Cardiovascular drugs | Angiotensin‐converting enzyme inhibitors | C09A | Included: Challenor 1991; Madsen 1984; |
| Cardiovascular drugs | Angiotensin‐II receptor antagonists | C09C | Included: none |
| Cardiovascular drugs | Nitrates | C01D | Included: Chung 2009; Nahir 1986; Teh 1995 |
| Unlisted | Peripheral vasodilators and related drugs | C04A | Included: Jaffe 1980; Le Quentrec 1991 |
| Psychotropic drugs | Selective serotonin reuptake inhibitors | N06A B | Included: none |
| Allergy and anaphylaxis | Antihistamines | N07CA | Included: none |
| Other (including Bradilan, Dazoxiben, UK‐38, 485 (dazmegrel)) | Included: Ettinger 1984 | ||
| AMH: Australian Medicines Handbook. aComparison between udenafil and amlodipine. | |||
| AMH | Class | ATC code | Drug name |
|---|---|---|---|
| Cardiovascular drugs | Vasodilators, antihypertensive drugs including | C02D, B01A | Included: beraprost sodium, cilostazol |
| Cardiovascular drugs | Phosphodiesterase inhibitors/phosphodiesterase type 5 inhibitors | G04BE | Included: vardenafil Ongoing: vardenafil |
| Cardiovascular drugs/endocrine drugs | Alpha adrenoreceptor‐blocking drugs | C02C | Included: thymoxamine |
| Cardiovascular drugs | Angiotensin‐converting enzyme inhibitors | C09A | Included: captopril, enalapril |
| Cardiovascular drugs | Angiotensin‐II receptor antagonists | C09C | Included: none |
| Cardiovascular drugs | Nitrates | C01D | Included: nitroglycerin/glyceryl trinitrate |
| Unlisted | Peripheral vasodilators and related drugs | C04A | Included: buflomedil, thymoxamine |
| Psychotropic drugs | Selective serotonin reuptake inhibitors | N06AB | Included: ketanserin |
| Allergy and anaphylaxis | Antihistamines | N07CA | Included: none |
| Others | Included: dazoxiben | ||
| AMH: Australian Medicines Handbook. | |||
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 1.1 Frequency of attacks per week Show forest plot | 3 | Mean Difference (IV, Random, 95% CI) | 0.79 [0.43, 1.16] | |
| 1.1.1 Captopril | 2 | Mean Difference (IV, Random, 95% CI) | 0.72 [‐1.73, 3.16] | |
| 1.1.2 Enalapril | 1 | Mean Difference (IV, Random, 95% CI) | 0.80 [0.43, 1.17] | |
| 1.2 Severity of symptoms Show forest plot | 2 | Mean Difference (IV, Random, 95% CI) | ‐0.17 [‐4.66, 4.31] | |
| 1.2.1 Captopril | 1 | Mean Difference (IV, Random, 95% CI) | ‐1.30 [‐7.18, 4.58] | |
| 1.2.2 Enalapril | 1 | Mean Difference (IV, Random, 95% CI) | 1.40 [‐5.55, 8.35] | |
| 1.3 Duration of attacks Show forest plot | 1 | Mean Difference (IV, Fixed, 95% CI) | Totals not selected | |
| 1.3.1 Captopril | 1 | Mean Difference (IV, Fixed, 95% CI) | Totals not selected | |
| 1.4 Capillaroscopic flow/skin perfusion by ultrasound Show forest plot | 1 | Mean Difference (IV, Fixed, 95% CI) | Totals not selected | |
| 1.4.1 Captopril | 1 | Mean Difference (IV, Fixed, 95% CI) | Totals not selected | |
| 1.5 Subjective assessment of improvement (10‐cm visual analogue scale) Show forest plot | 1 | Mean Difference (IV, Fixed, 95% CI) | Totals not selected | |
| 1.5.1 Enalapril | 1 | Mean Difference (IV, Fixed, 95% CI) | Totals not selected | |
| 1.6 Adverse events Show forest plot | 3 | 88 | Risk Ratio (M‐H, Fixed, 95% CI) | 1.35 [0.67, 2.73] |
| 1.6.1 Captopril | 2 | 48 | Risk Ratio (M‐H, Fixed, 95% CI) | 5.00 [0.27, 92.62] |
| 1.6.2 Enalapril | 1 | 40 | Risk Ratio (M‐H, Fixed, 95% CI) | 1.12 [0.55, 2.32] |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 2.1 Frequency of attacks ‐ buflomedil Show forest plot | 1 | Mean Difference (IV, Fixed, 95% CI) | Totals not selected | |
| 2.2 Severity of symptoms ‐ buflomedil Show forest plot | 1 | Mean Difference (IV, Fixed, 95% CI) | Totals not selected | |
| 2.3 Adverse events ‐ buflomedil Show forest plot | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | Totals not selected | |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 3.1 Frequency of attacks per week Show forest plot | 1 | Mean Difference (IV, Fixed, 95% CI) | Totals not selected | |
| 3.2 Severity of attacks (1 to 4 scale) Show forest plot | 1 | Mean Difference (IV, Fixed, 95% CI) | Totals not selected | |
| 3.3 Disability (100‐mm visual analogue scale) Show forest plot | 1 | Mean Difference (IV, Fixed, 95% CI) | Totals not selected | |
| 3.4 Adverse events Show forest plot | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | Totals not selected | |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 4.1 Frequency of attacks per week Show forest plot | 1 | Mean Difference (IV, Fixed, 95% CI) | Totals not selected | |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 5.1 Frequency of attacks per week Show forest plot | 1 | Mean Difference (IV, Fixed, 95% CI) | Totals not selected | |
| 5.2 Severity score (frequency of attacks/d × duration of attacks) Show forest plot | 1 | Mean Difference (IV, Fixed, 95% CI) | Totals not selected | |
| 5.3 Duration of attacks per day (minutes) Show forest plot | 1 | Mean Difference (IV, Fixed, 95% CI) | Totals not selected | |
| 5.4 Capillaroscopic flow/skin perfusion by ultrasound Show forest plot | 1 | Mean Difference (IV, Fixed, 95% CI) | Totals not selected | |
| 5.5 Adverse events Show forest plot | 1 | Risk Ratio (M‐H, Fixed, 95% CI) | Totals not selected | |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 6.1 Frequency of attacks Show forest plot | 2 | Mean Difference (IV, Random, 95% CI) | ‐1.57 [‐4.31, 1.17] | |
| 6.2 Severity of symptoms Show forest plot | 1 | Mean Difference (IV, Fixed, 95% CI) | Totals not selected | |
| 6.3 Capillaroscopic flow/skin perfusion by ultrasound Show forest plot | 1 | Mean Difference (IV, Fixed, 95% CI) | Totals not selected | |
| 6.4 Raynaud Condition Score Show forest plot | 1 | Mean Difference (IV, Fixed, 95% CI) | Totals not selected | |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 7.1 Frequency of attacks per week Show forest plot | 2 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.05 [‐6.71, 6.61] | |
| 7.1.1 Cilostazol | 1 | Std. Mean Difference (IV, Random, 95% CI) | 3.50 [0.64, 6.36] | |
| 7.1.2 PF‐00489791 | 1 | Std. Mean Difference (IV, Random, 95% CI) | ‐3.30 [‐3.92, ‐2.68] | |
| 7.2 Severity of symptoms Show forest plot | 2 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.03 [‐1.04, 0.97] | |
| 7.2.1 Cilostazol | 1 | Std. Mean Difference (IV, Random, 95% CI) | 0.50 [0.09, 0.91] | |
| 7.2.2 PF‐00489791 | 1 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.53 [‐0.63, ‐0.43] | |
| 7.3 Duration of attacks Show forest plot | 1 | Mean Difference (IV, Fixed, 95% CI) | Totals not selected | |
| 7.4 Capillaroscopic flow/skin perfusion by ultrasound Show forest plot | 1 | Mean Difference (IV, Fixed, 95% CI) | Totals not selected | |
| 7.5 Adverse events Show forest plot | 1 | 156 | Risk Ratio (M‐H, Fixed, 95% CI) | 1.49 [1.10, 2.03] |
| 7.6 Raynaud Condition Score Show forest plot | 2 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.80 [‐1.74, 0.13] | |
| 7.6.1 Vardenafil | 1 | Std. Mean Difference (IV, Random, 95% CI) | ‐1.61 [‐3.09, ‐0.13] | |
| 7.6.2 PF‐00489791 | 1 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.52 [‐0.60, ‐0.44] | |
