Pro‐coagulant haemostatic factors for the prevention and treatment of bleeding in people without haemophilia

Jez Fabes; Susan J Brunskill; Nicola Curry; Carolyn Doree; Simon J Stanworth

doi:10.1002/14651858.CD010649.pub2

Factores hemostáticos procoagulantes para la prevención y el tratamiento de la hemorragia en pacientes sin hemofilia

Declaraciones de intereses de los autores

Versión publicada: 24 diciembre 2018 Historial de versiones

https://doi.org/10.1002/14651858.CD010649.pub2

Contraer todo Desplegar todo

Resumen

disponible en

Antecedentes

Algunos pacientes ingresados pueden estar en riesgo de presentar, o presentar, una hemorragia grave. En estos pacientes son frecuentes las anomalías en la formación de coágulos (coagulación), y el tratamiento tradicional ha sido la transfusión de componentes sanguíneos, ya sea para prevenir la hemorragia (profiláctico) o para tratarla (terapéutico). Existe un interés creciente en el uso de tratamientos individualizados con concentrados de factores hemostáticos (que frenan la hemorragia y mantienen la sangre en el interior de un vaso sanguíneo dañado) procoagulantes específicos en lugar de plasma.

Objetivos

Evaluar los efectos y la seguridad de los factores hemostáticos procoagulantes y los concentrados de factores en la prevención y el tratamiento de las hemorragias en los pacientes sin hemofilia.

Métodos de búsqueda

Se buscaron ensayos controlados aleatorios (ECA) en el Registro Cochrane Central de Ensayos Controlados (Cochrane Cochrane Central Register of Controlled Trials) (2018, número 3), MEDLINE (desde 1948), Embase (desde 1974), CINAHL (desde 1938), PubMed (publicaciones en preparación hasta 18 de abril de 2018), PROSPERO, Transfusion Evidence Library (desde 1950), LILACS (desde 1980), IndMED (desde 1985), KoreaMed (desde 1934), Web of Science Conference Proceedings Citation Index (desde 1990) y en las bases de datos de ensayos en curso hasta el 18 de abril de 2018.

Criterios de selección

Se incluyeron los ECA que compararon la administración intravenosa de un concentrado de factores hemostáticos procoagulantes con placebo, el mejor tratamiento actual o habitual u otro concentrado de factores hemostáticos procoagulantes para la prevención o el tratamiento de la hemorragia. No hubo restricción en los tipos de participantes. Se excluyeron los estudios sobre desmopresina, ácido tranexámico y ácido aminocaproico y el uso de factores hemostáticos procoagulantes para la sobreanticoagulación con vitamina K.

Obtención y análisis de los datos

Se utilizaron los procedimientos metodológicos estándar Cochrane.

Resultados principales

Se identificaron 31 ECA con 2392 participantes y 22 ensayos en curso. En 13 ECA terapéuticos se asignaron al azar a 1057 participantes (intervalo: 20 a 249 participantes) y en 18 ensayos profilácticos se asignaron al azar a 1335 participantes (intervalo: 20 a 479 participantes). El concentrado de factores hemostáticos procoagulantes fue el fibrinógeno en 23 ensayos, el factor XIII en siete y los concentrados del complejo protrombínico (CCP) en uno.

Diecisiete ensayos tuvieron financiamiento o apoyo de la industria, ocho no declararon el financiamiento o la fuente de financiamiento no estuvo clara y seis estudios indicaron que no tuvieron una fuente de financiamiento de la industria.

Certeza en la evidencia y sesgo de los estudios incluidos

La certeza de la evidencia, sobre la base de los criterios GRADE, varió de muy baja a alta en todos los resultados. La certeza de la mayoría de los resultados se consideró baja. El riesgo de sesgo fue una inquietud en muchos de los ECA; la metodología de asignación al azar fue incierta en 15 ECA, la ocultación de la asignación fue incierta en 14 y el riesgo de sesgo fue alto en cinco. El estado del cegamiento de los evaluadores de resultados fue incierto en 13 ECA y tuvo alto riesgo de sesgo en cinco, aunque la mayoría de los resultados de estos ensayos fueron objetivos y no fueron propensos a sesgo del observador. El personal del estudio a menudo no estuvo cegado o no se disponía de información suficiente para evaluar el nivel de cegamiento (el riesgo de sesgo fue incierto en cinco ECA y alto en siete).

Resultados primarios

La mortalidad por todas las causas se informó en 21 ECA, los eventos tromboembólicos arteriales en 22 y los venosos en 21.

Concentrado de fibrinógeno: ensayos profilácticos con comparador inactivo (nueve ECA)

En los ensayos, los contextos clínicos y el momento de la medición de los resultados fueron heterogéneos, de manera que no se agruparon los datos. En comparación con placebo, no hubo evidencia de que el concentrado de fibrinógeno profiláctico redujera la mortalidad por todas las causas (cuatro ECA; 248 participantes). En comparación con los comparadores inactivos hubo evidencia de calidad baja a moderada de que el concentrado de fibrinógeno profiláctico no aumentó el riesgo de complicaciones tromboembólicas arteriales o venosas (siete ECA; 398 participantes).

Concentrado de fibrinógeno: ensayos profilácticos con comparador activo (dos ECA)

No hubo muertes ni incidencia de eventos tromboembólicos en estos dos ECA (con 57 participantes).

Concentrado de fibrinógeno: ensayos terapéuticos con comparador inactivo (ocho ECA)

En los ensayos, los contextos quirúrgicos y el momento de medición de los resultados fueron heterogéneos, de manera que sólo se agruparon los datos de los subgrupos. En comparación con un comparador inactivo, no hubo evidencia (la calidad varió de baja a alta) de que el concentrado de fibrinógeno redujera la mortalidad por todas las causas en los participantes con hemorragia activa (siete ECA; 724 participantes). En comparación con los comparadores inactivos, no hubo evidencia de que la administración del concentrado de fibrinógeno durante la hemorragia activa aumente los eventos tromboembólicos arteriales (siete ECA; 607 participantes) ni venosos (seis ECA; 562 participantes) .

Concentrado de fibrinógeno: ensayos terapéuticos con comparador activo (cuatro ECA)

No se agruparon los datos de resultado, ya que no se midieron en puntos temporales comparables. En comparación con otros agentes procoagulantes activos, no hubo evidencia (calidad muy baja a moderada) de que el concentrado de fibrinógeno redujera la mortalidad por todas las causas en pacientes con hemorragia activa cuatro ECA; 220 participantes). No hubo evidencia de que el concentrado de fibrinógeno aumentara el riesgo de eventos tromboembólicos arteriales (tres ECA; 126 participantes) ni venosos (cuatro ECA; 220 participantes).

Factor XIII: ensayos profilácticos con comparador inactivo (seis ensayos)

En los ensayos, los contextos quirúrgicos y el momento de la medición para el análisis de los resultados fueron heterogéneos por lo que sólo se agruparon los datos de los subgrupos. En comparación con un comparador inactivo, no hubo evidencia de que el factor XIII profiláctico redujera la mortalidad por todas las causas (cinco ECA; 414 participantes). No hubo evidencia (calidad muy baja a baja) de una diferencia en la tasa de eventos arteriales o venosos entre el factor XIII y los comparadores inactivos (cuatro ensayos; 354 participantes).

Factor XIII: ensayos terapéuticos con comparador inactivo (un ensayo)

No hubo muertes ni incidencia de eventos tromboembólicos en este ensayo.

Concentrado de complejo protrombínico (CCP): ensayos profilácticos con comparador inactivo (un ensayo)

No hubo evidencia (calidad moderada) de que el CCP redujera la mortalidad por todas las causas (un ensayo; 78 participantes). No se informaron de complicaciones tromboembólicas en este ensayo.

Conclusiones de los autores

La escasez de evidencia de buena calidad comparable impide establecer conclusiones para la práctica clínica. Se requieren estudios de investigación adicionales para determinar el cociente riesgo‐beneficio de estas intervenciones. Se debería aumentar de forma significativa el tamaño de la muestra de los ECA futuros para detectar una reducción en la mortalidad o los eventos tromboembólicos en los grupos de tratamiento. Para mejorar la consistencia en el informe de los resultados, es fundamental desarrollar grupos de resultados centrales, lo que puede ayudar a tratar algunas de las limitaciones identificadas en esta revisión.

PICO

Population

Intervention

Comparison

Outcome

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

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

Resumen en términos sencillos

disponible en

Efectividad de los fármacos para favorecer la coagulación en la prevención y el tratamiento de la hemorragia en pacientes sin hemofilia

Pregunta de la revisión

¿Los fármacos que ayudan en la prevención y el tratamiento de la hemorragia reducen el riesgo de muerte, el riesgo de formación de un coágulo sanguíneo y disminuyen la cantidad de sangre perdida en los pacientes que no presentan hemofilia (un trastorno hemorrágico) y que presentan, o están en riesgo de presentar, una hemorragia?

Antecedentes

La coagulopatía, definida como el fracaso de la coagulación normal de la sangre, es frecuente en las enfermedades graves, los traumatismos y la cirugía mayor. Estos cuadros pueden empeorar el sangrado y causar la muerte. Para el tratamiento de las coagulopatías y las hemorragias se dispone de fármacos que son factores de coagulación de la sangre administrados por vía intravenosa. Sin embargo, no hay información suficiente acerca de la efectividad y la seguridad.

Esta revisión combina todos los datos disponibles sobre estos fármacos para evaluar la efectividad y la seguridad.

Características de los estudios

Se realizaron búsquedas de ensayos controlados aleatorios (ECA) en la bibliografía médica hasta el 18 de abril de 2018, ya que éstos proporcionan la evidencia más confiable. Se identificaron 31 ECA relevantes con resultados de 2392 participantes, que compararon estos fármacos con placebo (tratamiento inactivo) u otro fármaco o producto sanguíneo. Los ECA se centraron en tres tipos de factores que pueden mejorar la coagulación de la sangre: el fibrinógeno (un factor de la coagulación que aumenta la fuerza del coágulo), el factor de coagulación XIII (importante para mantener los coágulos unidos) y el concentrado del complejo protrombínico (una combinación de cuatro precursores de factores de la coagulación).

Los ensayos administraron estos fármacos antes de que se presentara una hemorragia (como profilaxis) o para tratar la existente (terapéuticamente). La mayoría de los ensayos se centraron en la cirugía, especialmente la cirugía cardíaca, los traumatismos y la hemorragia posparto.

Diecisiete de los ECA identificados tuvieron apoyo de la industria, en ocho las fuentes de financiamiento no estuvieron claras y seis señalaron que no tuvieron financiamiento de la industria.

Resultados clave

Ningún fármaco tuvo efectos sobre el riesgo de muerte, independientemente del contexto clínico o de la forma en la que se utilizó. Sin embargo, la certeza de los resultados es baja, y este hallazgo puede cambiar en el futuro cuando se publiquen estudios nuevos.

Ningún fármaco aumentó el riesgo de coágulos nocivos en venas o arterias, pero la certeza de estos resultados es baja.

El fibrinógeno profiláctico redujo la hemorragia después de la cirugía cardíaca y ortopédica en comparación con placebo. El fibrinógeno profiláctico, en comparación con placebo, redujo casi a la mitad la necesidad de una transfusión de sangre posterior a la cirugía cardíaca y redujo en el 75% la necesidad en otras cirugías. El fibrinógeno redujo la necesidad de una transfusión de sangre cuando se utilizó para el tratamiento de la hemorragia.

El factor profiláctico XIII redujo la hemorragia después de la cirugía cardíaca.

Se debe aumentar de forma significativa el tamaño de la muestra de los ensayos aleatorios futuros para poder mostrar cualquier diferencia en la supervivencia general y la muerte por hemorragia.

Certeza de la evidencia

La certeza de la evidencia es baja, pero los estudios de investigación futuros pueden cambiar los hallazgos de esta revisión.

Conclusión

La certeza baja de la evidencia dificulta establecer conclusiones acerca de la efectividad de estos fármacos y si se deben utilizar en la asistencia sanitaria actual. Se necesitan estudios de investigación adicionales en la forma de ECA de gran tamaño para determinar los costes de estos tratamientos, los efectos beneficiosos y si éstos superan los riesgos.

Authors' conclusions

Implications for practice

The inadequate quality of evidence in most of the studies included in this review means that we cannot draw conclusions for clinical practice or the use of these interventions outside controlled trials. Further research is required in larger‐scale randomised controlled trials to determine the risk‐to‐benefit ratio of these interventions. We anticipate that data from the identified ongoing trials, some of which are larger studies, will provide much higher‐quality evidence around clinical outcomes and efficacy.

At present we lack a clear picture of the cost effectiveness of these interventions, due to a lack of health economic data. It is likely that many pro‐haemostatic agents are more costly than blood components, including plasma. Most studies co‐administer the pro‐coagulant agents with a broad range of other haemostatic interventions as part of the clinical setting or routine management of haemorrhage. Additionally, the progression towards viscoelastic testing and algorithmic clotting factor approaches to transfusion means that these agents are often administered in combination. However, at present we lack the data required to assess how these interventions interact or work together.

Implications for research

The main implication for this review is to direct the need for further research. Whilst at one level the numbers of trials represents a body of data to inform clinical practice, the limitations preclude the drawing of robust conclusions without additional studies. There is considerable current clinical interest in the use of these agents (we identified 22 ongoing trials), which is likely to expand in the coming years. Sample sizes in future randomised trials powered for clinical outcomes of mortality (or even preferably mortality due to bleeding) would need to be greatly increased. Analogous examples for power calculations would be CRASH‐2 (CRASH‐2 2010), WOMAN (WOMAN 2017) and HALT IT (Brenner 2018; Roberts 2014), which recruited between 15,000 and 20,000 participants each.

To improve consistency in outcome reporting, the use of core outcome sets (COS) are increasingly advocated by both the CONSORT and SPIRIT statements (Moher 2010; Chan 2013), and may help address a number of the limitations in reporting identified in this review. A COS is defined as a “minimum (number of outcomes) that should be measured and reported in all clinical trials of a specific condition” (Gargon 2017). The Core Outcome Measures in Effectiveness Trials (COMET) database has assisted in driving international collaboration amongst clinical researchers across a plethora of clinical conditions, such as critical care and rheumatological diseases, and it may be timely to look at this methodology for outcomes of bleeding. In particular, consideration of how to combine clinically meaningful endpoints (such as achievement of haemostasis) with reliable laboratory data must be a focus for future studies, in order to improve the quality of the available evidence for laboratory thresholds to guide clinicians about when to administer haemostatic therapy. Future updates of this review may need to be split by interventions used as treatment versus prophylaxis.

Researchers may also consider the lessons from previous research in the management of coagulopathy and bleeding. Firstly, the extensive trial programmes undertaken to evaluate recombinant factor VIIa consisted initially of poor‐quality studies with small participant numbers in a heterogeneous range of clinical settings which demonstrated variable degrees of clinical efficacy that overestimated the benefit while concealing the underlying harm. It is essential that this process is not repeated. Secondly, there were two large studies of tranexamic acid (TXA) that demonstrated a positive impact on death from bleeding. CRASH‐2 was a study of just over 20,000 participants that showed that the administration of TXA reduced overall mortality and death from bleeding at a statistically significant level (CRASH‐2 2010). Subsequent analysis of the data, in combination with data from over 20,000 participants in the WOMAN study of TXA in postpartum haemorrhage (WOMAN 2017), found that administration of an early dose of TXA improved outcomes. The multicentre and multi‐national design, the pragmatic approach and large participant numbers of these studies permitted statistically robust and clinically useful conclusions to be drawn about practice in a real‐world setting.

Summary of findings

Open in table viewer

Summary of findings for the main comparison. FIBRINOGEN: Therapeutic trials: intervention compared to inactive Control for the prevention and treatment of bleeding in participants without haemophilia

FIBRINOGEN: Therapeutic trials: intervention compared to inactive control for the prevention and treatment of bleeding in participants without haemophilia
Patient or population: the prevention and treatment of bleeding in patients without haemophilia Setting: hospital Intervention: FIBRINOGEN in therapeutic trials Comparison: Inactive control
Outcomes	*Anticipated absolute effects^ (95% CI)**		Relative effect (95% CI)	№ of participants (studies)	Certainty of the evidence (GRADE)	Comments
	Risk with inactive control	Risk with fibrinogen in therapeutic trials
All‐cause mortality Up to 28 days following admission	Study population		RR 1.46 (0.71 to 2.99)	97 (2 RCTs)	⊕⊕⊕⊝ MODERATE^a	‐
	184 per 1000	268 per 1000 (130 to 549)
All‐cause mortality In‐hospital mortality up to 30 days post‐operatively	Study population		RR 5.00 (0.25 to 102.00)	120 (1 RCT)	⊕⊕⊝⊝ LOW^a,b	‐
	0 per 1000	0 per 1000 (0 to 0)
All‐cause mortality Up to 6 weeks post‐natally	Study population		‐	294 (2 RCTs)	⊕⊕⊝⊝ LOW^a,b	No events
	see comment	see comment
All‐cause mortality Up to 46 days post‐operatively	Study population		RR 0.23 (0.05 to 1.01)	213 (2 RCTs)	⊕⊕⊕⊕ HIGH	‐
	85 per 1000	20 per 1000 (4 to 86)
*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). CI: Confidence interval; RR: Risk ratio
GRADE Working Group grades of evidence High certainty: We are very confident that the true effect lies close to that of the estimate of the effect Moderate certainty: We are moderately confident in the effect estimate: The true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different Low certainty: Our confidence in the effect estimate is limited: The true effect may be substantially different from the estimate of the effect Very low certainty: We have very little confidence in the effect estimate: The true effect is likely to be substantially different from the estimate of effect
^aWe downgraded by one level as neither study was designed to measure mortality. ^bWe downgraded by one level, due to low event rate.

Open in table viewer

Summary of findings 2. FIBRINOGEN: Therapeutic trials: intervention compared to haemostatically active control for the prevention and treatment of bleeding in participants without haemophilia

FIBRINOGEN: Therapeutic trials: intervention compared to haemostatically active control for the prevention and treatment of bleeding in participants without haemophilia
Patient or population: the prevention and treatment of bleeding in patients without haemophilia Setting: hospital Intervention: FIBRINOGEN in therapeutic trials Comparison: Haemostatically active control
Outcomes	*Anticipated absolute effects^ (95% CI)**		Relative effect (95% CI)	№ of participants (studies)	Certainty of the evidence (GRADE)	Comments
	Risk with haemostatically active control	Risk with FIBRINOGEN in therapeutic trials
All‐cause mortality Up to 7 days post‐operatively or hospital discharge	Study population		not estimable	63 (1 RCT)	⊕⊕⊕⊝ MODERATE^a	No events
	0 per 1000	0 per 1000 (0 to 0)
All‐cause mortality Up to 28 days post‐operatively	Study population		not estimable	20 (1 RCT)	⊕⊕⊝⊝ LOW^a,b	No events
	0 per 1000	0 per 1000 (0 to 0)
All‐cause mortality Up to 30 days post‐operatively	Study population		RR 0.95 (0.06 to 14.30)	43 (1 RCT)	⊕⊕⊝⊝ LOW^a,c	‐
	48 per 1000	45 per 1000 (3 to 681)
All‐cause mortality Up to 30 days post‐operatively	Study population		RR 2.20 (0.45 to 10.78)	94 (1 RCT)	⊕⊝⊝⊝ VERY LOW^a,c,d	‐
	45 per 1000	100 per 1000 (20 to 490)
*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). CI: Confidence interval; RR: Risk ratio
GRADE Working Group grades of evidence High certainty: We are very confident that the true effect lies close to that of the estimate of the effect Moderate certainty: We are moderately confident in the effect estimate: The true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different Low certainty: Our confidence in the effect estimate is limited: The true effect may be substantially different from the estimate of the effect Very low certainty: We have very little confidence in the effect estimate: The true effect is likely to be substantially different from the estimate of effect
^aWe downgraded by one level due to low event rate. ^bWe downgraded by one level due to concerns about risks of bias both in the way that allocation of treatment was concealed at randomisation and whether blinding to treatment could be maintained for the duration of the trial in this "open‐label trial where adequacy of surgical field haemostasis reported as a subjective outcome". ^cWe downgraded by one level as the study did not measure mortality as an outcome. ^dWe downgraded by one level due to concerns about risks of bias in how blinding to treatment intervention could be maintained for the duration of the trial in this "open‐label trial where adequacy of surgical field haemostasis reported as a subjective outcome", and the non‐reporting of all outcomes that the investigators said they would measure (selective outcome reporting).

Open in table viewer

Summary of findings 3. FIBRINOGEN: Prophylactic trials: intervention compared to inactive control for the prevention and treatment of bleeding in participants without haemophilia

FIBRINOGEN: Prophylactic trials: intervention compared to inactive control for the prevention and treatment of bleeding in participants without haemophilia
Patient or population: the prevention and treatment of bleeding in patients without haemophilia Setting: hospital Intervention: FIBRINOGEN: prophylactic trials Comparison: inactive control
Outcomes	*Anticipated absolute effects^ (95% CI)**		Relative effect (95% CI)	№ of participants (studies)	Certainty of the evidence (GRADE)	Comments
	Risk with inactive control	Risk with FIBRINOGEN in prophylactic trials
Venous thromboembolic events Up to 30 days post‐operatively	Study population		‐	122 (2 RCTs)	⊕⊕⊝⊝ LOW^a,b	No events in one trial
	see comment	see comment
Arterial thrombotic events Up to 4 days post‐operatively	Study population		not estimable	80 (2 RCTs)	⊕⊕⊕⊝ MODERATE^a	‐
	25 per 1000	0 per 1000 (0 to 0)
Arterial thrombotic events Up to 30 days post‐operatively	Study population		RR 0.33 (0.01 to 8.02)	116 (1 RCT)	⊕⊕⊝⊝ LOW^a,c	‐
	17 per 1000	6 per 1000 (0 to 138)
Arterial thrombotic events ‐ up to 30 days post‐operatively	Study population		not estimable	122 (2 RCTs)	⊕⊕⊝⊝ LOW^a,b	‐
	51 per 1000	0 per 1000 (0 to 0)
Venous thrombotic events Up to 4 days post‐operatively	Study population		not estimable	80 (2 RCTs)	⊕⊕⊕⊝ MODERATE^a	‐
	0 per 1000	0 per 1000 (0 to 0)
*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). CI: Confidence interval; RR: Risk ratio
GRADE Working Group grades of evidence High certainty: We are very confident that the true effect lies close to that of the estimate of the effect Moderate certainty: We are moderately confident in the effect estimate: The true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different Low certainty: Our confidence in the effect estimate is limited: The true effect may be substantially different from the estimate of the effect Very low certainty: We have very little confidence in the effect estimate: The true effect is likely to be substantially different from the estimate of effect
^aWe downgraded by one level due to low event rate. ^bWe downgraded by one level due to concerns about risks of bias in the non‐reporting of all outcomes that the investigators said they would measure (selective outcome reporting). ^cWe downgraded by one level due to concerns about risks of bias in the way that allocation of the treatment was concealed at the time of randomisation.

Open in table viewer

Summary of findings 4. FIBRINOGEN: Prophylactic trials: intervention compared to haemostatically active control for the prevention and treatment of bleeding in participants without haemophilia

FIBRINOGEN: Prophylactic trials: intervention compared to haemostatically active control for the prevention and treatment of bleeding in participants without haemophilia
Patient or population: the prevention and treatment of bleeding in patients without haemophilia Setting: hospital Intervention: FIBRINOGEN: prophylactic trials Comparison: Haemostatically active control
Outcomes	*Anticipated absolute effects^ (95% CI)**		Relative effect (95% CI)	№ of participants (studies)	Certainty of the evidence (GRADE)	Comments
	Risk with haemostatically active control	Risk with FIBRINOGEN in prophylactic trials
Arterial thromboembolic events	Study population		not estimable	49 (1 RCT)	⊕⊕⊝⊝ LOW^a,b	‐
	0 per 1000	0 per 1000 (0 to 0)
Arterial thromboembolic events ‐ Duration of hospital stay (whole population)	Study population		not estimable	49 (1 RCT)	⊕⊕⊝⊝ LOW^a,b	‐
	0 per 1000	0 per 1000 (0 to 0)
Venous thromboembolic events	Study population		not estimable	49 (1 RCT)	⊕⊕⊝⊝ LOW^a,b	‐
	0 per 1000	0 per 1000 (0 to 0)
Venous thromboembolic events Duration of hospital stay (whole population)	Study population		not estimable	49 (1 RCT)	⊕⊕⊝⊝ LOW^a,b	‐
	0 per 1000	0 per 1000 (0 to 0)
*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). CI: Confidence interval; RR: Risk ratio
GRADE Working Group grades of evidence High certainty: We are very confident that the true effect lies close to that of the estimate of the effect Moderate certainty: We are moderately confident in the effect estimate: The true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different Low certainty: Our confidence in the effect estimate is limited: The true effect may be substantially different from the estimate of the effect Very low certainty: We have very little confidence in the effect estimate: The true effect is likely to be substantially different from the estimate of effect
^aWe downgraded by one level as the thromboembolic events (arterial or venous) in these participants given fibrinogen prophylactically were not an outcome of interest. ^bWe downgraded by one level due to the small number of participants available for this outcome.

Open in table viewer

Summary of findings 5. FXIII: Prophylactic trials: intervention compared to inactive control for the prevention and treatment of bleeding in participants without haemophilia

FXIII: Prophylactic trials: intervention compared to inactive control for the prevention and treatment of bleeding in participants without haemophilia
Patient or population: the prevention and treatment of bleeding in patients without haemophilia Setting: hospital Intervention: FXIII: prophylactic trials Comparison: inactive control
Outcomes	*Anticipated absolute effects^ (95% CI)**		Relative effect (95% CI)	№ of participants (studies)	Certainty of the evidence (GRADE)	Comments
	Risk with inactive control	Risk with FXIII in prophylactic trials
Arterial thromboembolic events Up to 5 to 7 weeks post‐operatively	Study population		RR 0.67 (0.28 to 1.62)	282 (2 RCTs)	⊕⊕⊝⊝ LOW^a,b	‐
	81 per 1000	54 per 1000 (23 to 131)
Arterial thromboembolic events Up to 30 days post‐operatively	Study population		RR 3.00 (0.14 to 66.53)	22 (1 RCT)	⊕⊕⊝⊝ LOW^b,c	‐
	0 per 1000	0 per 1000 (0 to 0)
Arterial thromboembolic events Post‐operatively	Study population		RR 0.20 (0.01 to 3.97)	50 (1 RCT)	⊕⊝⊝⊝ VERY LOW^b,c,d	‐
	80 per 1000	16 per 1000 (1 to 318)
Venous thromboembolic events ‐ Up to 5 to 7 weeks post‐operatively	Study population		RR 0.93 (0.06 to 14.67)	282 (2 RCTs)	⊕⊕⊝⊝ LOW^a,b	‐
	7 per 1000	7 per 1000 (0 to 108)
Venous thromboembolic events Up to 30 days post‐operatively	Study population		RR 3.00 (0.14 to 66.53)	22 (1 RCT)	⊕⊕⊝⊝ LOW^b,c	‐
	0 per 1000	0 per 1000 (0 to 0)
Venous thromboembolic events Post‐operatively	Study population		not estimable	50 (1 RCT)	⊕⊝⊝⊝ VERY LOW^b,c,d	‐
	0 per 1000	0 per 1000 (0 to 0)
*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). CI: Confidence interval; RR: Risk ratio
GRADE Working Group grades of evidence High certainty: We are very confident that the true effect lies close to that of the estimate of the effect Moderate certainty: We are moderately confident in the effect estimate: The true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different Low certainty: Our confidence in the effect estimate is limited: The true effect may be substantially different from the estimate of the effect Very low certainty: We have very little confidence in the effect estimate: The true effect is likely to be substantially different from the estimate of effect
^aWe downgraded by one level due to concerns about attrition, reporting and selection bias across these trials. ^bWe downgraded by one level as thromboembolic events (arterial or venous) in these participants given factor XIII prophylactically were not an outcome of interest. ^cWe downgraded by one level due to low event rates. ^dWe downgraded by one level due to risks of bias concerns in how concealment of treatment intervention was attempted (unsealed envelopes in the patients notes) and the subsequent impact this had on the level of blinding to treatment intervention that could be maintained for the duration of the trial.

Background

A term which we use frequently throughout this review is 'haemostatic'. Haemostatic means to cause bleeding to stop, keeping blood within a damaged blood vessel or tissue.

Description of the condition

Coagulopathy (a dysfunctional clotting response) is common in critically‐ill people, both in those with and without bleeding, because of the nature of their illness. Coagulopathy is typically defined by abnormalities of standard coagulation tests performed in haematology laboratories, most commonly defined by prolongation of the prothrombin time or the activated partial thromboplastin time. These are in vitro tests of the time taken for a plasma sample to start to form a fibrin clot. Notably, the prothrombin time or the activated partial thromboplastin time or both are poor predictors of bleeding and there are no clear definitions of a threshold that defines a clinically significant prolongation of either test (Tripodi 2009).

In a recent, multi‐centre, prospective, observational trial in adult critical care, which evaluated just under 2000 intensive‐treatment‐unit (ITU) admissions, one‐third of patients without active bleeding were shown to develop coagulopathy during their ITU admission (Walsh 2009).

The most common current treatment for people with coagulopathy remains the transfusion of plasma. Plasma is the liquid component of blood in which different cellular elements and macromolecules are suspended. Plasma contains pro‐coagulant haemostatic proteins which are a key component of the haemostatic system. These pro‐coagulant haemostatic proteins, known as blood clotting factors, interact with each other, with platelets and tissues to form stable blood clots. Around one‐third of people with coagulopathy received transfusions of plasma (termed fresh frozen plasma (FFP) in the UK), with around half of all FFP transfusions given to reduce a perceived risk of bleeding or prior to invasive procedures (Stanworth 2011; Walsh 2009).

However, there is uncertainty around the effectiveness of plasma transfusion in clinical practice, and the risk‐benefit profile of FFP remains unknown (Yang 2012). The recognised risks associated with its use include fluid overload, transfusion‐related acute lung injury, along with the generic risks associated with transfusion of allogeneic blood products (SHOT 2011). A national audit of 5000 FFP transfusion episodes to consecutive patients from 196 hospitals across England indicated little efficacy for plasma transfusion, with very small median change in duration of prothrombin time following FFP transfusion (Stanworth 2010).

Description of the intervention

Given concerns about the effectiveness of plasma transfusion, there is increasing interest in the targeted use of specific or purified intravenous pro‐coagulant haemostatic factors. Many different intravenous pro‐coagulant haemostatic factors exist, including those derived from plasma as concentrates, or those developed as specific recombinant factors. The advantages of plasma‐derived coagulation factor concentrates include greater standardisation (whether at the stage of production or during manufacture) and reduced infective and compatibility complications when compared with blood transfusions through the application of techniques such as pasteurisation, nanofiltration and adsorption to remove viruses. The advantages of recombinant‐derived factors include potentially unlimited supply, with reduced reliance on the availability of blood donors, as well as negligible infective risk. Finally, both recombinant and plasma‐derived intravenous concentrates offer the potential advantage of providing specific pro‐coagulant haemostatic factors more rapidly and at higher concentrations than is possible with the use of traditional plasma transfusion.

Some of the plasma‐derived coagulation factor concentrates contain multiple pro‐coagulant haemostatic factors. Prothrombin complex concentrates are plasma‐derived coagulation factor concentrates that contain three or four (vitamin K‐dependent) factors at high concentration.

Two observational cohort studies have examined the use of prothrombin complex concentrates as an adjunct to fibrinogen concentrate in the management of bleeding following trauma. Whilst neither of these studies demonstrated reduced mortality in those receiving prothrombin complex concentrates, one of the studies demonstrated an association between receiving prothrombin complex concentrates and a reduced transfusion requirement (Görlinger 2011). Safety data on the use of prothrombin complex concentrates are limited. In particular, rates of thrombosis, both arterial and venous, and disseminated intravascular coagulation need to be fully assessed (Grottke 2011).

In contrast to concentrates containing multiple factors, single pro‐coagulant haemostatic factors allow the treatment of individual clotting factor deficiencies to be specifically targeted. A specific concentrate of the key pro‐coagulant haemostatic factor fibrinogen (or factor I) is available, although currently only licensed for use in many countries for people with inherited disorders of fibrinogen. The observation that normal‐to‐high levels of fibrinogen (in the range 1.5 to 4.0 g/L) prior to the onset of major obstetric haemorrhage are associated with improved outcomes (Charbit 2007; de Lloyd 2011), has led clinicians to consider targeted therapy with concentrates of fibrinogen in this cohort of women, rather than the more traditional use of FFP.

How the intervention might work

Specific coagulation factor concentrates are presumed to stop or prevent bleeding by providing a source of pro‐coagulant haemostatic factor(s) which, alongside contributions from platelets and the endothelium, interact to produce a stronger fibrin clot. The rationale for their use is based on either restoring or increasing levels of pro‐coagulant haemostatic factors, or activating clotting systems by infusing recombinant activated complexes to initiate coagulation.

It is hypothesised that pro‐coagulant haemostatic factors administered intravenously might work more effectively than plasma (either FFP or cryoprecipitate) to deliver a more concentrated source of factors which will raise factor levels more effectively and predictably using a standardised product. Given that the current processes for thawing plasma would not apply to pro‐coagulant haemostatic factors, it is also likely that these agents would be more rapidly available for use in emergency situations.

However, as for all pro‐coagulant haemostatic agents, adverse events are recognised. In particular, use of these agents might be expected to be associated with a greater rate of thromboembolic events. This has been clearly demonstrated in meta‐analyses of RCTs of recombinant factor VIIa in people without haemophilia (Levi 2010; Simpson 2012).

Why it is important to do this review

In a short period there has been a rapid shift in the clinician's approach to the management of coagulopathy and bleeding, with increasing interest in the use of specific intravenous pro‐coagulant haemostatic factors instead of the current standard plasma transfusion. Because of the real uncertainties about the effectiveness and safety of FFP, wider use of the pro‐coagulant haemostatic factors and evaluation in clinical trials are anticipated, not least to understand their benefits given their increased cost in comparison to plasma transfusion. There is increasing understanding of the complex changes that occur in people with acquired major bleeding, particularly after trauma and parturition, and it is clear that many clotting factor levels change during active bleeding.

This review explores both the use of single clotting factor concentrate therapy (at the time of publication these were fibrinogen, factor XIII and a prothrombin complex concentrate composite) but we also anticipate that clotting factor concentrates will be evaluated in combination to treat the complexity of clotting changes that occurs in major haemorrhage. Lessons should also be learned from the story of rFVIIa, which was 'marketed' as a universal haemostatic agent in people without haemophilia but was then found, as new randomised trials were published, to have uncertain benefit with harm (increased arterial thromboembolic events; Levi 2010; Simpson 2012), emphasising the importance of a systematic review to identify and appraise the RCT literature as it evolves. This Cochrane Review builds on the experience gained from the rFVIIa Cochrane Review (Simpson 2012), and comprehensively identifies and critically appraises all the RCT evidence assessing the use of intravenous pro‐coagulant haemostatic factors.

This systematic review differs from two other 'clotting factor concentrate' Cochrane Reviews (Wikkelso 2013; Johansen 2015). Our review did not assess the use of pro‐coagulant haemostatic factors and factor concentrates for the reversal of vitamin K antagonist treatment in bleeding and non‐bleeding people, which is the focus of Johansen 2015. Our review does not include trials where the intervention was recombinant factor VIIa, antifibrinolytic therapy or desmopressin, as is the case in Wikkelso 2013. Our review includes trials in which the participant group were neonates, unlike Wikkelso 2013 which excluded them.

This systematic review indicates where high‐quality evidence is lacking, so informing healthcare professionals where more evidence is required before these expensive drugs are widely accepted into clinical practice.

Objectives

To assess the effects and safety of pro‐coagulant haemostatic factors and factor concentrates in the prevention and treatment of bleeding in people without haemophilia.

Methods

Criteria for considering studies for this review

Types of studies

We included only randomised controlled trials (RCTs), regardless of language, date or place of publication. We included open‐label and unblinded studies.

Types of participants

We included adults and children (with no age restriction) either at risk of bleeding or already bleeding due to any cause. We excluded people with known inherited bleeding disorders (e.g. haemophilia) unless any data were separately available for the participants without the inherited condition. We excluded studies of healthy volunteers.

Types of interventions

We assessed separately whether a pro‐coagulant haemostatic factor concentrate was administered as a treatment or in a prophylactic manner for the following comparisons:

a pro‐coagulant haemostatic factor concentrate (either alone or in combination with another pro‐coagulant haemostatic factor concentrate) versus either placebo or current best/standard treatment (e.g. plasma/FFP).
one pro‐coagulant haemostatic factor concentrate (either alone or in combination with another pro‐coagulant haemostatic factor concentrate) versus another different pro‐coagulant haemostatic factor concentrate (either alone or in combination with another pro‐coagulant haemostatic factor concentrate).

Each of the above combinations could be given in any dose and for any duration of treatment.

We included RCTs of the administration of prothrombin complex concentrates if used for conditions other than reversal of vitamin K antagonist therapy.

We only included studies where pro‐coagulant haemostatic factors were administered intravenously.

We excluded:

pro‐coagulant haemostatic factors administered topically, e.g. thrombin or fibrin glue, as they are the subject of a previous Cochrane Review (Carless 2009).
RCTs of desmopressin, tranexamic acid and aminocaproic acid, due to their different mechanisms of action from other pro‐coagulant haemostatic agents. We also excluded RCTs in which a pro‐coagulant haemostatic agent was used exclusively to reverse warfarin (or a similar agent), but after the review had been started.
Studies exclusively administering rFVIIa, as they have been the subject of a previous Cochrane Review (Simpson 2012). However, studies in which rFVIIa was a co‐intervention with other intravenous pro‐coagulant haemostatic factors (i.e. in separate comparator arms or co‐administration of rFVIIa and another factor/reagent) were eligible for inclusion.

We are aware that studies of fibrinogen concentrate in bleeding have been the focus of a previous Cochrane Review (Wikkelso 2013), but as this is a rapidly expanding field with a significant number of new studies published since the Wikkelso 2015 review, and as we have presented the analysis from an alternative perspective, we have included the fibrinogen concentrate studies in this review,

Types of outcome measures

We presented outcome data from all identified RCTs by prophylactic and therapeutic indications for use. In prophylactic use, the agent was given to prevent bleeding which may be anticipated prior to or during a surgical operation; in therapeutic use, the agent was given to treat bleeding which is established and ongoing.

There was variation in the time points at which each outcome was measured, due to the nature of the participants and conditions across the trials; we took this into account when reporting and analysing the outcome data for the included trials. Most of the trials reported outcomes acutely postpartum and in the 24 hours following any given intervention or procedure, when compared to baseline measures for outcomes such as blood loss and transfusion requirement. These formed our main time point comparators for these outcomes. In addition,where outcome data were reported for other time points we have also included these data. For mortality, arterial and venous thromboembolic events, bleeding and adverse events, data were presented at different time points in each trial, making it difficult to generate a cohesive picture.

Primary outcomes

Mortality (all causes) (principal primary outcome for therapeutic intervention usage)
The number of participants suffering arterial and/or venous thromboembolic events (principal primary outcome for prophylactic intervention usage)

Secondary outcomes

Mortality due to bleeding
Red blood cell transfusion requirement (both volume of transfusion and the receipt of a transfusion)
Blood loss
Allergic reaction
Change to laboratory measures of coagulation, both standard laboratory tests and global measurements of coagulation (such as thromboelastography (TEG) or rotational thromboelastometry (ROTEM), or thrombin generation)

Search methods for identification of studies

The Systematic Review Initative's Information Specialist (CD) formulated the search strategies used, in collaboration with the Cochrane Injuries Group.

Electronic searches

We searched for RCTs in the following electronic databases between the dates listed below. There were no restrictions by language or publication status. All searches were undertaken on 18 April 2018 and the search strategies for all sources are detailed in Appendix 1.

Cochrane Central Register of Controlled Studies (CENTRAL) (the Cochrane Library 2018, Issue 3) on 18 April 2018;
MEDLINE (OvidSP, 1948 to 18 April 2018);
Embase (OvidSP, 1974 to 18 April 2018);
CINAHL (EBSCOHOST, 1938 to 18 April 2018);
PubMed, in process and epublications ahead of print only, on 18 April 2018;
PROSPERO on 18 April 2018;
Transfusion Evidence Library (www.transfusionevidencelibrary.com, 1950 to 18 April 2018);
LILACS (from 1980 to 18 April 2018);
IndMed (from 1985 to 18 April 2018);
KoreaMed (from 1934 to 18 April 2018);
Web of Science Conference Proceedings Citation Index ‐ Science (CPCI‐S, from 1990 to 18 April 2018).

We also searched for ongoing trials in the following databases. We undertook all searches on 18 April 2018 and the search strategies are detailed in Appendix 1.

ClinicalTrials.gov (www.clinicaltrials.gov);
EU Clinical Trials Register (EUDRACT, www.clinicaltrialsregister.eu);
WHO International Clinical Trials Registry Platform (ICTRP, www.who.int/ictrp)
ISRCTN Register (www.isrctn.com/)

We combined searches in MEDLINE, Embase and CINAHL with adaptations of the Cochrane RCT search filter, as detailed in Chapter 6 of theCochrane Handbook for Systematic Reviews of interventions (Higgins 2011).

Searching other resources

We augmented database searching by:

Handsearching of reference lists; we checked references of all identified trials and relevant review articles for further literature. We limited these searches to 'first generation' reference lists;
Personal contacts; we contacted authors of relevant trials, trial groups and experts worldwide who are known to be active in the field for unpublished material or further information on ongoing trials.

Data collection and analysis

Selection of studies

One author (CD) initially screened all search results for relevance against the eligibility criteria and discarded all those that were clearly irrelevant. Thereafter, three review authors (JF, NC, YL) independently screened all the remaining results (titles, abstracts and full text) for relevance against the full eligibility criteria. We retrieved full‐text papers for all references for which a decision of eligibility was not made from the title and abstract alone. We resolved differences of opinion through discussion and consensus, where necessary, with recourse to a fourth review author (SB). Trials that did not meet our eligibility criteria are detailed in the Characteristics of excluded studies table.

Data extraction and management

Five review authors (JF, NC, SB, YL, GS) independently extracted data onto standardised forms. Two of the review authors (JF, GS) piloted these forms on two included RCTs and made changes that were agreed upon as needed. Throughout the data extraction process, we resolved disagreements by consensus, with recourse to a third review author (SS) as needed. The review authors were not blinded to the names of study authors, institutions, journals or the outcomes of the trials. We extracted data on the setting of the trial, the methods and statistical assumptions made, the characteristics of participants and interventions and details of the outcomes measured and the timing of the outcome assessments.

Where studies describe red blood cell transfusion in units (Jeppsson 2016; Karlsson 2011; Lance 2012; Najafi 2014; Sadeghi 2014), we converted these to millilitres to permit comparison with other studies reporting millilitres of red blood cells transfused. We have used a conversion ratio of one unit of red blood cells equivalent to 300 millilitres.

Assessment of risk of bias in included studies

Three review authors (JF, SB, NC) independently assessed all included trials for possible risks of bias, as described in Chapter 8 of the Cochrane Handbook (Higgins 2011). We classified trials as being at low, high or unclear risk of bias. To assess risks of bias, the authors included the following questions in the 'Risk of bias' table for each included trial.

Was the allocation sequence adequately generated?
Was allocation adequately concealed?
Was knowledge of the allocated intervention adequately prevented (i.e. blinded) throughout the trial for participants, trial personnel and outcome assessors?
Were incomplete outcome data adequately addressed for the main outcome?
Are reports of the trial free of selective outcome reporting?
Was the trial apparently free of other problems that could put it at risk of bias?

We anticipated that knowledge of the allocated intervention would be difficult to blind from clinicians and participants. However we felt that it would be possible to blind knowledge of the allocated intervention from outcome assessors. We bore these considerations in mind when making an assessment of the risk of bias for the adequacy of blinding of participants, trial personnel and outcome assessors.

Measures of treatment effect

We present dichotomous outcomes as relative risks (RRs) with 95% confidence intervals (CIs). We present continuous outcomes as mean differences (MDs) with 95% confidence intervals. Where data are not presented in this way, but we were able to make adjustments from the data given, we did this for the following studies:

Godje 2006; we present the data in tabular form without quoting numeric values. We extracted these data quantitatively using digital measurement of the relevant table data and axes.

A number of studies reported outcome data for blood loss and transfusion requirement using median values with interquartile ranges (IQRs) (12 and 17 studies, respectively); we have commented on these values in the text and tabulated them in Table 1 and Table 2.

Open in table viewer

Table 1. Red blood cell transfusion requirement

Trial	Number included in analyses intervention / comparator	Clinical condition	Transfusion Protocol?	When was transfusion requirement measured	How were findings reported	Findings: Intervention/ Comparator
Fibrinogen concentrate: prophylactic trials with inactive comparator
Cui 2010	17/14	Surgery (E) : cardiac	Not stated	During ICU stay	Median (IQR) number of red blood cell units transfused	0 (0 ‐ 0)/0 (0 ‐ 1.25)
					Median (IQR) total red blood cell usage	0 (0 ‐ 1.5)/1.5 (0 ‐ 2.3)
Fenger Erikson 2009	10/10	Surgery (E): radical cystectomy	Not stated	During surgery	Median (IQR) number of red blood cell units transfused	2 (0 ‐ 5) / 2.5 (0 ‐ 6)
				48 hours post‐operatively		0 (0 ‐ 2)/1.5 (0 ‐ 2)
				Total transfused		3.5 (0 ‐ 5)/4.0 (0 ‐ 6)
				Within 48 hours post‐operatively	Total number (%) patients requiring a red blood cell transfusion	2 (20%)/8 (80%)
Karlsson 2011	10/10	Surgery (E) : cardiac	Yes	Within 12 hours post‐operatively	Mean (SD) number of red blood cell units transfused	0.2 (0.13)/0.7 (0.27)
Najafi 2014	15/15	Surgery (E): total hip replacement arthroplasty	Yes	"Perioperative period"	Mean (SD) "sum of transfused packed cells"	0.8 (1.01)/1.06 (1.2)
Pournajafian 2015	21/20	Surgery (E): posterior spinal fusion	Not stated	Up to 24 hours post‐operatively	Number (%) patients transfused	0 (0%)/6 (30%)
Ranucci 2015	58/58	Surgery (E): cardiac	Yes	Duration of hospital stay until 30 days	Total number (%) of patients receiving a red blood cell transfusion	19 (33%)/32 (55%)
					Median (IQR) number of units transfused	0 (0 ‐ 1)/1 (0 ‐ 2)
				Within 24 hours of surgery	Mean (SD) mL of red blood cells infused for transfused patients only	136 (178)/356 (488)
				From 24 to 48 hours after surgery	Mean (SD) mL of red blood cells infused for transfused patients only	420 (760)/218 (552)
Sabate 2016	48/44	Surgery (E): liver transplantation	Yes	During surgery	Median (IQR) number of units transfused	0 (0 ‐ 1)/0 (0 ‐ 0.5)
					Total number (%) of patients receiving a red blood cell transfusion	17 (52.9%)/11 (42.7%)
				During and up to 24 hours after surgery	Median (IQR) number of units transfused	2 (0 ‐ 6)/3 (0 ‐ 6)
					Total number (%) of patients receiving a red blood cell transfusion	33 (68.8%)/30 (68.2%)
Sadeghi 2014	30/30	Surgery (E) : cardiac	Yes	Within first 24 hours	Mean (SD) number of red blood cell infusions	1.5 (1.8)/2 (1.5)
					Total number (%) of patients receiving a red blood cell transfusion	15 (50)/21 (70)
Soleimani 2016	31/29	Surgery (E): transurethral resection of the prostate.	Not stated	post‐operatively	Total number (%) of patients receiving a red blood cell transfusion	1 (3.2%)/5 (17.2%)
Fibrinogen concentrate: prophylactic trials with active comparator
Haas 2015	10/9	Surgery (E): scoliosis	Yes	Over 24 hours from the start of surgery	Median (IQR) mL/kg of red blood cells transfused	0 (0 ‐ 15.3)/18.4 (0 ‐ 23.8)
					Mean (SD) mL/kg of red blood cells transfused	3.7 (6.2)/7.2 (8.0)
					Total number (%) of patients receiving a red blood cell transfusion	4 (40%)/6 (67%)
	17/ 13	Surgery (E): craniosynostosis	Yes	Over 24 hours from the start of surgery	Median (IQR) mL/kg of red blood cells transfused	28.2 (21.2‐ 49.9)/55.5 (27.5 ‐ 61.8)
					Mean (SD) mL/kg of red blood cells transfused	33.1 (17.9)/41.8 (18.1)
					Total number (%) of patients receiving a red blood cell transfusion	17 (100%)/13 (100%)
Fibrinogen concentrate: therapeutic trials with inactive comparator
Bilecen 2017	58/59	Surgery (E): cardiac	Yes	Between study intervention and chest closure	Total number (%) of patients receiving a red blood cell transfusion	0 (0%)/3 (5%)
				Between study intervention and 24 hours	Total number (%) of patients receiving a red blood cell transfusion	10 (17%)/20 (33%)
					Median (IQR) number of units transfused	0 (0 ‐ 1)/0 (0 ‐ 4)
Collins 2017	28/27	postpartum haemorrhage	Not stated	From study intervention to discharge	Median (IQR) number of units transfused	1 (0 – 2)/1 (0 – 2)
					Total number (%) of patients who received a red blood cell transfusion	13 (46.6%)/13 (48.1%)
					Total number of red blood cell transfusions	37/38
Curry 2018	24/24	Trauma	Yes	3hrs following admission	Median (IQR) number of units transfused	4 (2 ‐ 6)/2 (2 ‐ 6)
					Total number (%) of patients receiving a red blood cell transfusion	21 (91.3%)/20 (87.0%)
				6hrs following admission	Median (IQR) number of units transfused	3 (2 ‐ 6)/2 (2 ‐ 6)
					Total number (%) of patients receiving a red blood cell transfusion	19 (90.5%)/19 (86.4%)
				24hrs following admission	Median (IQR) number of units transfused	4 (2 ‐ 8)/2 (2 ‐ 5)
					Total number (%) of patients receiving a red blood cell transfusion	19 (95.0%)/18 (85.7%)
Jeppsson 2016	24/24	Surgery (E): cardiac	Yes	During hospital stay	Mean (SD) units of red blood cell s transfused	0.63 (1.17)/1.33 (3.11)
					Median (IQR) number of units transfused	0 (0 ‐ 1)/0 (0 ‐ 1)
Nascimento 2016	21/24	Trauma	Yes	At 24 hours	Median (IQR) number of units transfused	3 (2 ‐ 5)/3 (2 ‐ 4)
Rahe‐Meyer 2013	29/32	Surgery (E): aortic valve replacement	Yes	24 hours after receipt of study medication	Total number (%) of patients receiving a red blood cell transfusion	16 (55)/32 (100)
					Median (IQR) number of units transfused	0 (0 ‐ 3)/2 (2‐5)
Rahe‐Meyer 2016	78/74	Surgery (E): complex cardiac	Yes	24 hours after receipt of study medication	Median (IQR) number of units transfused	1 (0 ‐ 3)/0 (0 ‐ 2)
Wikkelso 2015	123/121	postpartum haemorrhage	Yes	Need for a red blood cell transfusion up to 4 hours post study drug	Total number (%) of patients receiving a red blood cell transfusion	4 (3)/10 (8)
				Need for a red blood cell transfusion up to 24 hours post study drug		14 (11)/19 (16)
				Need for a red blood cell transfusion up to 7 days post study drug		25 (20)/26 (21)
				"During the 6 week period postpartum"		25 (20)/26 (21)
				Time‐point not stated	Total number (%) of patients who received a transfusion of > 4 units of red blood cells.	8 (7)/3 (3)
Fibrinogen concentrate: therapeutic trials with active comparator
Galas 2014	30/33	Surgery (E): cardiac	Yes	post‐operative day 7	Total number (%) of patients receiving a red blood cell transfusion	25 ( 83)/32 (97)
Tanaka 2014	10/10	Surgery (E): valve replacement or repair	Yes	At 24 hours post‐operatively	Total number (%) of patients receiving a red blood cell transfusion	9 (90)/9 (90)
Lance 2012	22/21	Surgery (E): cardiovascular, abdominal or spinal column	Yes	Following study intervention	Mean (SD) number of red blood cell units transfused	4.98 (3.01)/5.38 (2.38)
Innerhofer 2017	50/44	Trauma	Yes	At 24 hours after injury	Total number (%) of patients receiving a red blood cell transfusion	25 (90%)/39 (89%)
					Median (IQR) number of red blood cell units transfused	4 (2 ‐ 7)/6 (4 ‐ 11)
Factor XIII (FXIII): prophylactic trials with inactive comparator
Godje 2006	25/25	Surgery (E) : cardiac	Yes	At 24 hours post‐operatively	Mean (SD) number of red blood cell units transfused	2.6 (1.875)/3.3 (1.397)
Karkouti 2013	138/128	Surgery (E) : cardiac	Not stated	Until post‐operative day 7 or hospital discharge (whichever occurred first)	Median (IQR) number of red blood cell units transfused	0 (0‐1)/0 (0‐1)
					Total number (%) of patients receiving a red blood cell transfusion	46 (33)/43 (34)
Korte 2009	22 in total	Surgery (E): gastrointestinal cancer	Yes	Unclear time point	Median (IQR) mL of red blood cells transfused	150 (0‐700)/200 (0‐800)
Levy 2009	8/8	Surgery (E) : cardiac	Not stated	Until post‐operative day 7 or hospital discharge (whichever occurred first)	Total number of patients requiring allogeneic blood products	No data was reported
					Percentage of patients avoiding a red blood cell transfusion	No data was reported
Rasche 1982	29/31	Acute leukaemia	Yes	Not stated	Median (IQR) number of red blood cell units transfused	8 (0 ‐ 34)/6 (0 ‐ 18)
Factor XIII (FXIII): therapeutic trials with inactive comparator
PCC: prophylactic trials with inactive comparator

(E) = elective surgery; ICU = intensive care unit, IQR = inter quartile range; SD = standard deviation.

Open in table viewer

Table 2. Blood loss

Trial	Number included in analyses intervention/comparator	Clinical condition	Transfusion Protocol?	How the outcome was measured	Findings: intervention	Findings: Comparator
Fibrinogen concentrate: prophylactic trials with inactive comparator
Cui 2010	17/14	Surgery (E) : cardiac	Yes	Mean (SD) quantity (mL/kg/h) of chest drainage at 1 h post‐operatively	2.9 (2.0) mL/kg/h	3.5 (1.6) mL/kg/h
				Mean (SD) quantity (ml/kg/h) of chest drainage at 6 h post‐operatively	1.3 (0.1) mL/kg/h	1.5 (0.6) mL/kg/h
				Median (IQR) quantity ( mL/kg/h) of chest drainage at 24 h post‐operatively	0.6 (0.4 to 0.8) mL/kg/h	0.7 (0.6 to 0.9) mL/kg/h
Fenger Erikson 2009	10/10	Surgery (E): radical cystectomy	Not stated	Mean (SD) quantity (mL) blood loss at time of intervention	2682 (962) mL	2933 (1320) mL
Karlsson 2011	10/10	Surgery (E) : cardiac	Yes	Mean (SD): "post‐operative bleeding"	565 (150) mL/12 h	830 (268) mL/12 h
Najafi 2014	15/15	Surgery (E): total hip replacement arthroplasty	Yes	Mean (SD) quantity (mL) of intraoperative blood loss	744.6 (337) mL	870 (498) mL
Najafi 2014	15/15	Surgery (E): total hip replacement arthroplasty	Yes	Mean (SD) quantity (mL) of post‐ operative bleeding to 24 hours post surgery	231.6 (67) mL	236.6 (81) mL
Pournajafian 2015	21/20	Surgery (E): posterior spinal fusion	Not stated	Mean (SD) mL blood loss up to 24 hours post‐operatively	533.3 (157.9) mL	679 (130) mL
Ranucci 2015	58/58	Surgery (E): cardiac	Yes	Median (IQR) quantity (mL) lost from chest drain in first 12 hours	300 (200 to 400) mL/12 h	355 (250 to 600) mL/12 h
Sadeghi 2014	30/30	Surgery (E) : cardiac	Yes	Mean (SD) quantity (mL) of blood lost in chest drain in first 12 h post‐operatively	477 (143) mL	703 (179) mL
Soleimani 2016	31/29	Surgery (E): transurethral resection of the prostate.	Not stated	Mean (SD) quantity (mL) of blood lost intraoperatively	521 (290) mL	557 (411) mL
Soleimani 2016	31/29	Surgery (E): transurethral resection of the prostate.	Not stated	Mean (SD) quantity (mL) of blood lost post‐operatively	291 (270) mL	341 (314) mL
Fibrinogen concentrate: prophylactic trials with active comparator
Haas 2015	10/9	Surgery (E): scoliosis	Yes	Median (IQR) "calculated total blood loss (%)" by 24 h post‐operatively	36.5 (14.9 to 54.3) %	51.0 (38.5 to 69.2) %
Haas 2015	17/ 13	Surgery (E): craniosynostosis	Yes	Median (IQR) "calculated total blood loss (%)" by 24 h post‐operatively	89.7 (77.7 to 112.9) %	156.9 (110..5 to 187.3) %
Fibrinogen concentrate: therapeutic trials with inactive comparator
Bilecen 2017	58/59	Surgery (E): cardiac	Yes	Median (IQR) mL blood loss from study intervention to chest closure	50 (29 ‐ 100) mL	70 (33 ‐ 145) mL
				Median (IQR) mL blood loss from ICU arrival to 1hr	70 (35 ‐ 130) mL	90 (46 ‐ 149) mL
				Median (IQR) mL blood loss in ICU from 1hr to 3h	80 (50 ‐ 156) mL	110 (40 ‐ 220) mL
				Median (IQR) mL blood loss in ICU from 3h to 6h	100 (54 ‐ 169) mL	110 (60 ‐ 208) mL
				Median (IQR) mL blood loss in ICU from 6h to 12h	110 (80 ‐ 160) mL	125 (83 ‐ 224) mL
				Median (IQR) mL blood loss in ICU from 12h to 24h	130 (80 ‐ 180) mL	160 (90 ‐ 270) mL
				Median (IQR) cumulative mL blood loss in 24 hours	570 (390 ‐ 730) mL	690 (400 ‐ 1090) mL
Collins 2017	28/27	postpartum haemorrhage	Not stated	Median (IQR) mL blood loss within 24h of study intervention	225 (100 – 341) mL	300 (60 – 800) mL
Jeppsson 2016	24/24	Surgery (E): cardiac	Yes	Mean (SD) mL intraoperative blood loss	306 (110) mL	375 (202) mL
				Median (IQR) mL intraoperative blood loss	300 (200–400) mL	300 (200–500) mL
				Mean (SD) mL post‐operative blood loss (within 12h)	796 (523) mL	897 (553) mL
				Median (IQR) mL post‐operative blood loss (within 12h)	650 (500–835) mL	730 (543–980) mL
				Mean (SD) mL total blood loss (within 12h)	1103 (518) mL	1272 (588) mL
				Median (IQR) mL total blood loss (within 12h)	913 (815–1230) mL	1185 (930–1398) mL
Rahe‐Meyer 2016	78/74	Surgery (E): complex cardiac	Yes	Median (IQR) mL chest tube drainage volume at 6h	260.0 (155.0–410.0) mL	297.5 (200.0–455.0) mL
				Median (IQR) mL chest tube drainage volume at 12h	405.0 (245.0–600.0) mL	447.5 (320.0–700.0) mL
				Median (IQR) mL chest tube drainage volume at 24h	590.0 (400.5–839.5) mL	682.5 (530.0–1050.0) mL
Wikkelso 2015	123/121	postpartum haemorrhage	Yes	Median (IQR) quantity (mL) lost to 6 weeks postpartum	1700 (1500 to 2000) mL	1700 (1400 to 2000) mL
Fibrinogen concentrate: therapeutic trials with active comparator
Galas 2014	30/33	Surgery (E): cardiac	Yes	Median (IQR) quantity (mL) chest drainage from intraoperative to 48 h post‐operatively	320 (157 ‐ 750) mL	410 (215 ‐ 510) mL
Tanaka 2014	10/10	Surgery (E): valve replacement or repair	Yes	Median (IQR) quantity (mL) chest tube drainage at 12 h post‐operatively	925 (500 to 1693) mL	1315 (653 to 2965) mL
Factor XIII (FXIII): prophylactic trials with inactive comparator
Godje 2006	25/25	Surgery (E) : cardiac	Yes	Mean (SD) drain volume (ml) at 6 h post‐operatively^$	231 (113) mL	278 (98) mL
				Mean (SD) drain volume (ml) at 12 h post‐operatively^$	371 (170) mL	458 (160) mL
				Mean (SD) drain volume (ml) at 24 h post‐operatively^$	649 (227) mL	747 (222) mL
				Mean (SD) drain volume (ml) at 36 h post‐operatively^$	798 (273) mL	969 (309) mL
				Mean (SD) drain volume (ml) at 48h post‐operatively^$	958 (376) mL	1231 (175) mL
Korte 2009	22 in total	Surgery (E): gastrointestinal cancer	Yes	Median (IQR) quantity (ml) blood loss at completion of surgery	750 (400 to 1000) mL	1050 (700 to 1800) mL
Levy 2009	8/8	Surgery (E) : cardiac	Not stated	Mean (SD) quantity (mL) of chest drainage at 8 h post‐operatively	358 (93.5) mL	505 (301) mL
				Mean (SD) quantity (mL) of chest drainage at 24h post‐operatively	903 (738) mL	1155 (804) mL
				Mean (SD) quantity (mL) of chest drainage at CT removal	1007 (912) mL	1489 (1395) mL
Factor XIII (FXIII): therapeutic trials with inactive comparator
PCC: prophylactic trials with inactive comparator

^ "Calculated total blood loss (%) was based on the estimated total blood volume (calculated as described by Kearney 1989 ) " Haas 2015

* "The visual assessment of the surgical field was performed by the senior surgical staff as follows: 0 = excellent haemostasis (dry field), 1 = mild bleeding (oozing), 2 = moderate bleeding (controllable with applied pressure) and 3 = severe bleeding (multiple diffuse bleeding sites)" Tanaka 2014

$ Findings are reported graphically, not numerically.

(E) = elective surgery; h = hours; ICU = intensive care unit, IQR = inter quartile range; mL = millilitres; SD = standard deviation.

Unit of analysis issues

In the included trials the participant is the unit of analysis.

We did not encounter unit‐of‐analysis issues. However, if in any future updates of this review we identify any cluster‐randomised trials we will treat these in accordance with the advice given in the Cochrane Handbook (Higgins 2011).

Dealing with missing data

We contacted study authors by email to request missing data or to clarify data that were presented. The list of authors we contacted and the information obtained is outlined below.

Collins 2017 provided additional data about mortality.
Curry 2018 provided additional data about arterial and venous thromboembolic adverse events and the number of participants exposed to a red blood cell transfusion.
Haas 2015 (C) and Haas 2015 (S) provided additional data about the volume of red blood cell units transfused, the number of participants exposed to a red blood cell transfusion, mortality, bleeding mortality and arteriovenous thromboembolic events.
Karkouti 2013 provided additional data on adverse events.
Korte 2009 provided additional information on the number of participants in each study arm.
Lance 2012 provided additional data about outcome time points, mortality, adverse events and red blood cell transfusion.
Ranucci 2015 provided additional information about the number of participants exposed to a red blood cell transfusion.
Sabate 2016 provided additional data on the incidence of bleeding mortality, arterial and venous thromboembolic events and the volume of red blood cell units transfused.
Innerhofer 2017, Jeppsson 2016, Karlsson 2011 and Wikkelso 2015 were unable to provide any additional data that we could include in this review.
Bilecen 2017; Cui 2010; Galas 2014; Levy 2009; Najafi 2014; Nascimento 2016; Pournajafian 2015; Rahe‐Meyer 2013; Rahe‐Meyer 2016; Sadeghi 2014; Soleimani 2016; and Tanaka 2014 provided no response to our communication.
The contact email addresses of Bregenzer 1999; Fenger Erikson 2009; Godje 2006 generated an automatic failure‐to‐deliver email communication and no alternative contact details were available.
There were no relevant or usable contact details provided in Rasche 1982; Shirahata 1990; Turner 1981, and no alternative contact details were available.

Method of analysis

Of the 31 included trials, 11 declared an intention‐to‐treat (ITT) analysis, 13 outlined per protocol analysis methods and seven studies did not identify their method of analysis. Due to this, and the paucity of data that could be usefully harmonised on participant cohort, intervention and outcome time point we have included per protocol data.

Assessment of heterogeneity

We assessed clinical heterogeneity based on the data extracted about the characteristics of the included trials. We have discussed possible explanations for the observed heterogeneity within the review.

Where we considered trials to be clinically homogeneous, we conducted meta‐analysis and assessed statistical heterogeneity of treatment effects between trials by using a Chi² test with a significance level at P < 0.1. We used the I² index to quantify the amount of possible heterogeneity (I² > 30% as moderate heterogeneity and I² > 75% as considerable heterogeneity). When we found heterogeneity, we explored its potential causes with prespecified subgroup and sensitivity analyses.

Assessment of reporting biases

We made every effort to identify unpublished trials relevant to this review. We accept that funnel‐plot asymmetry, of which publication bias is one cause, is difficult to detect with the small number of trials (i.e. less than 10) often encountered in systematic reviews. In future updates, we will assess publication bias using funnel plots if we include 10 or more trials in any of our predefined comparison subgroups.

Data synthesis

We undertook meta‐analysis where trials were sufficiently homogeneous in terms of clinical setting and trial design and there were sufficient data of a suitable type. We used the Cochrane statistical software Review Manager 5 (Review Manager 2014). We used a fixed‐effect model for combining data in the first instance. Where we identified heterogeneity in a fixed‐effect meta‐analysis, we noted this and repeated the analysis using a random‐effects model. We did not remove any study from any analysis due to a high risk of bias. We grouped outcome data into the outcomes listed earlier in the Types of outcome measures section.

As well as quantitative synthesis, the overall interpretation of the data incorporated insights from a qualitative summary. We based our conclusions on patterns of results identified across clearly‐tabulated results of included trials, as well as summary measures. We considered both the direction and magnitude of any effect. We document individual trial results in the 'Additional Tables' section of the review.

For meta‐analysis, we log‐transformed blood loss outcome data and conducted the analysis using the transformed values. A meta‐analysis (using arithmetic means) of the differences in means using the transformed blood loss data corresponds to a meta‐analysis (using geometric means) of the ratio of means in the original scale.

Subgroup analysis and investigation of heterogeneity

We performed one subgroup analysis due to heterogeneity, using the first of our predefined subgroups.

Children versus adults (because different definitions of coagulopathy apply). We defined adults as being 18 years or older at the time of recruitment.
Trials of participants with brain/neurological injuries. For example, acute intracranial injury versus no acute intracranial injury (in view of the added effect of central nervous system injury on coagulopathy and therefore the possibly different efficacy of interventions in this participant group).

We did not use the second of our predefined subgroups as these data were sparse in the included studies.

Sensitivity analysis

We assessed the robustness of our findings by performing sensitivity analyses for the trials where 1) allocation concealment was adequate, and 2) in which 25% of participants or fewer were lost in the one meta‐analysis with noted heterogeneity.

'Summary of findings' tables

We created 'Summary of findings' tables for each intervention, following guidance in the Cochrane Handbook (Higgins 2011). We assessed the quality of the evidence for each outcome using GRADE (Schünemann 2011). The outcomes included in the 'Summary of findings' tables are:

All‐cause mortality for trials in which the indication of use for the intervention was therapeutic;
Arterial and venous thromboembolic complications for trials in which the indication of use for the intervention was prophylactic.

Results

Description of studies

Results of the search

The search strategy identified 4781 references with an additional 379 ongoing studies; these were reduced to 2770 references plus 319 ongoing studies once duplicates had been removed. Of these, 2681 references and 288 ongoing studies were ruled out for being clearly irrelevant to the scope of this review. We screened the remaining 89 references and 31 ongoing trials for eligibility. We included 31 studies (from 61 references) and identified 22 ongoing trials (from 25 references); we excluded 14 trials (from 28 references) and six ongoing studies (from six references). Full details are reported in the PRISMA flow diagram (Figure 1).

Figure 1

Trial flow diagram.

We found four non‐English publications. Two papers required translation from German to assess eligibility for inclusion (Rasche 1982; Scherer 1994). Following translation, we deemed one paper ineligible (Scherer 1994), whilst we rated the second trial as eligible for inclusion (Rasche 1982). A third paper was translated from Portuguese using Google translate and we included it in the review (Sadeghi 2014). The fourth paper was translated from Persian by a volunteer translator found through the Cochrane TaskExchange (Pournajafian 2015), and was also included.

One reference reported data from two distinct surgical groups: scoliosis surgery and craniosynostosis surgery (Haas 2015). We report these data as two separate studies throughout the review: Haas 2015 (S) and Haas 2015 (C).

The 31 included studies randomised a total of 2392 participants. There were 13 therapeutic trials that randomized 1057 participants (ranging from 20 to 249) and 18 prophylactic trials that randomised a total of 1335 participants (ranging from 20 to 479). Two trials had three arms (Godje 2006; Karkouti 2013), and one trial had five arms (Levy 2009). All three were prophylactic trials.

Included studies

We give further details about the included studies in the Characteristics of included studies table and Table 3; Table 4; Table 5; Table 6; Table 7; Table 8 and Table 9.

Open in table viewer

Table 3. Fibrinogen concentrate: prophylactic trials with inactive comparator

Trial	Number randomised (N = intervention/N = comparator)	Number included in analyses (N = intervention/N = comparator)	Intervention	Comparator	Co‐interventions	Primary Outcomes	Clinical Condition	Population Age
Cui 2010	40 (20/20)	31 (17/14)	Fibrinogen concentrate (0.5‐1g) combined with traditional transfusion, guided by TEG	NONE: Traditional transfusion guided by clinical experience.	Protamine, FFP, red blood cells, platelets.	Time to chest wall closure.	Surgery (E) : cardiac	Paediatric population
Fenger Erikson 2009	21 (10/11 )	20 (10/10)	Fibrinogen concentrate (45mg/kg), administered intra‐operatively	PLACEBO: isotonic saline, administered intra‐operatively at a dose of 2.25 mL/kg).	Intra‐operative FFP, red blood cells.	Whole blood maximum clot firmness as determined by thromboelastometry.	Surgery (E): radical Cystectomy	Adult population
Karlsson 2011	20 (10/10)	20 (10/10)	Preoperative infusion offibrinogen concentrate (2g) at 5 minutes following baseline measurements.	NONE	Heparin, protamine, red blood cells, aspirin.	Clinical adverse events, including graft occlusion on CT up to 3 to 4d post‐ operatively.	Surgery (E) : cardiac	Adult population
Najafi 2014	30 (15/15)	30 (15/15)	Fibrinogen concentrate (30 mg/kg) infused after induction of general anaesthesia	PLACEBO: normal saline administered in equal volume as fibrinogen.	None stated.	Volume of red blood cells transfused during and 24h post‐operatively.	Surgery (E): total hip replacement arthroplasty	Adult population
Pournajafian 2015	41 (21/20)	41 (21/20)	Fibrinogen concentrate (1g) infused after induction of general anaesthesia.	PLACEBO: normal saline.	None stated.	Not stated.	Surgery (E): posterior spinal fusion.	Adult population
Ranucci 2015	116 (58/58)	116 (58/58	Fibrinogen concentrate with dose determined by the mean clot firmness test on the FIBTEM.	PLACEBO: Normal (0.9%) saline.	TXA during surgery, protamine.	Avoidance of allogeneic blood products transfusion during hospital stay up to 30d.	Surgery (E): cardiac	Adult population
Sabate 2016	99 (51/48)	92 (48/44)	Fibrinogen concentrate with dose calculated to reach a target plasma concentration of 2.9 g/L.	PLACEBO: Normal (0.9%) saline.	Red blood cells, platelets, FFP, tranexamic acid.	The percentage of patients requiring transfusion of red blood cell units during the liver transplant procedure.	Surgery (E): liver transplantation.	Adult population
Sadeghi 2014	60 (30/30)	60 (30/30)	Fibrinogen (1g) administered 30 minutes prior to anaesthesia induction.	PLACEBO: saline administered 30 minutes prior to anaesthesia induction	None stated.	Volume of post‐operative haemorrhage 0, 12 and 24h post‐operatively.	Surgery (E): cardiac	Adult population
Soleimani 2016	72 (36/36)	60 (31/29)	Fibrinogen concentrate (2g) administered 15‐30 minutes before the start of surgery.	PLACEBO: 50ml normal saline.	None stated.	Bleeding volume during and after surgery.	Surgery (E): transurethral resection of the prostate.	Adult population

* For the outcomes bleeding and transfusion requirements, 1 patient from the Factor XIII, 1250 IE arm and 2 patients from the placebo arm were excluded from the analysis due to re‐operations following hints of early bypass occlusions. All patients were included in the analysis of the other outcomes.

d = days; (E) = elective surgery; FIBTEM = is a ready‐to‐use ROTEM^® system reagent for use with citrated whole blood and assesses the clot firmness of the fibrin clot, h = hours; PCC = prothrombin complex concentrate; mins = minutes; wks = weeks.

Open in table viewer

Table 4. Fibrinogen concentrate: prophylactic trials with active comparator

Trial

Number randomised

(N = intervention/ N = comparator)

Number included in analyses

(N = intervention/ N = comparator)

Intervention

Comparator

Co‐interventions

Primary Outcomes

Clinical Condition

Population Age

Haas 2015

Craniosynostosis surgery = 36

(17/14).

Scoliosis surgery = 26

(10/9)

Craniosynostosis surgery = 36

(17/13).

Scoliosis surgery = 26

(10/9)

Fibrinogen concentrate if FIBTEM MCF under 13mm (early substitution group).

Total dose administered per subject in

craniosynostosis surgery; median of 90mg kg^‐1 (IQR, 75 to 120 mg kg^‐1);

Scoliosis surgery = = median of 60mg kg^‐1 (IQR, 30 to 68 mg kg^‐1).

Fibrinogen concentrate if FIBTEM MCF under 8 mm (conventional).

Total dose administered per subject in

craniosynostosis surgery = median of

90mg kg^‐1 (IQR, 60 to 90 mg kg^‐1); Scoliosis surgery = median of 30mg kg^‐1 (IQR, 30 to 60 mg kg^‐1).

Tranexamic acid

Total volume of transfused red blood cells per kg bodyweight within 24h post‐operatively

Surgery (E): craniosynostosis or scoliosis

Paediatric population

CPB = cardiopulmonary bypass; (E) = elective surgery; FFP = fresh frozen plasma; FIBTEM = is a ready‐to‐use ROTEM^® system reagent for use with citrated whole blood and assesses the clot firmness of the fibrin clot; h = hours; INR = International Normalised Ratio, IQR = inter quartile range; PCC = prothrombin complex concentrate; (U) = urgent surgery/ treatment.

Open in table viewer

Table 5. Fibrinogen concentrate: therapeutic trials with inactive comparator

Trial	Number randomised (N = intervention/ N = comparator)	Number included in analyses (N = intervention/ N = comparator)	Intervention	Comparator	Co‐interventions	Primary Outcomes	Clinical Condition	Population Age
Bilecen 2017	120 (60/60)	115 (58/57)	Fibrinogen concentrate, dose calculated to provide a target plasma fibrinogen concentration of 2.5 g/L.	PLACEBO: 2g of albumin diluted with 50ml 0.9% sodium chloride solution.	Red blood cells, platelets, FFP, tranexamic acid.	The intraoperative blood loss measured between intervention and closure of the chest.	Surgery (E): cardiac	Adult population
Collins 2017	57 (29/28)	55 (28/27)	Fibrinogen concentrate, dose determined by FIBTEM to elevate A5 result above 22mm.	PLACEBO: 50ml of normal saline.	Red blood cells, platelets, FFP, tranexamic acid.	The number of allogeneic blood products (red blood cell , FFP, cryoprecipitate, platelets) infused after study medication until hospital discharge.	postpartum haemorrhage	Adult population
Curry 2018	48 (24/24)	48 (24/24)	Fibrinogen concentrate 6g infusion.	PLACEBO: 300ml of normal saline.	Major haemorrhage therapy as per hospital protocol: red blood cells, platelets, FFP, tranexamic acid, cryoprecipitate.	Feasibility of delivery of FgC therapy within 45 minutes to adult trauma patients and the proportion of participants whose fibrinogen levels were maintained ≥ 2 g/L during active haemorrhage.	Trauma	Adult population
Jeppsson 2016	52 (26/26)	48 (24/24)	Fibrinogen (2g) as an infusion over 5 minutes, immediately before surgery.	PLACEBO: 100ml of normal saline.	300 units kg−1 of heparin, 1 mg protamine per 100 units of heparin. Red blood cells, platelets, FFP, additional fibrinogen.	Mediastinal drain loss during the first 12 hours post‐operatively.	Surgery (E): cardiac	Adult population
Nascimento 2016	50 (25/25)	45 (21/24)	Fibrinogen concentrate, 6g as a 3 minute infusion.	PLACEBO: 300ml of normal saline as a 3 minute infusion.	Red blood cells, tranexamic acid, FFP, platelets and cryoprecipitate.	The feasibility of providing the study intervention within 1 hour of hospital admission.	Trauma	Adult population
Rahe‐Meyer 2013	80 (38 /42)	71 (29/32)	Fibrinogen concentrate, dose determined by FIBTEM test: median (IQR range) dose: 8g (6 to 9g).	PLACEBO: normal (0.9%) saline. Median (IQR range) dose: 400ml (300 to 450ml).	TXA during surgery, platelets or FFP.	Total number of units of allogeneic blood components given during first 24h post‐operatively	Surgery (E): aortic valve replacement	Adult population
Rahe‐Meyer 2016	152 (78/74)	152 (78/74)	Fibrinogen concentrate, dose determined by FIBTEM testing to generate an MCF of 22mm.	PLACEBO: 0.9% sodium chloride solution.	Red blood cells, platelets, FFP, tranexamic acid.	The number of units of allogeneic blood products (FFP, platelets and red blood cells) administered during the 24 hours after administration of study medication.	Surgery (E): complex cardiac	Adult population
Wikkelso 2015	249 (124 /125)	249 (PPA: 123 /121) (ITT: 120 /119)	Fixed dose of 2g of fibrinogen concentrate dispensed using a syringe pump infusion over 20 minutes	PLACEBO: 100ml of isotonic saline (mean (standard deviation) dose =1.2 (0.2) mg/kg).	TXA, hydroxyethyl starch, intravenous fluids.	Red blood cell transfusion during a 6 wk follow‐up period postpartum	postpartum haemorrhage	Adult population

d = days; (E) = elective surgery; FFP = fresh frozen plasma; FIBTEM = is a ready‐to‐use ROTEM^® system reagent for use with citrated whole blood and assesses the clot firmness of the fibrin clot; h = hours; INR = International Normalised Ratio, IQR = inter quartile range; mins = minutes; PCC = prothrombin complex concentrate; TXA = tranexamic acid; (U) = urgent surgery/ treatment; wks = weeks.

Open in table viewer

Table 6. Fibrinogen concentrate: therapeutic trials with active comparator

Trial	Number randomised (N = intervention/ N = comparator)	Number included in analyses (N = intervention/ N = comparator)	Intervention	Comparator	Co‐interventions	Primary Outcomes	Clinical Condition	Population Age
Galas 2014	63 (30/33)	63 (30/33)	Single dose of Fibrinogen concentrate (60mg/kg body weight, Haemocomplettan®) given at time of intra‐operative bleeding.	Single dose of Cryoprecipitate (10ml/kg body weight) given at time of intra‐operative bleeding.	Not Stated	post‐operative blood losses during 48h after surgery	Surgery (E): cardiac	Paediatric population
Lance 2012	43 (22/21)	43 (22/21) (but only 16/16 for primary outcome analysis).	FFP ( 2 units) + Fibrinogen (2g)	FFP (4 units)	Protamine, red blood cells	Pre‐ and post‐transfusion ROTEM analysis	Surgery (E): cardiovascular, abdominal or spinal column	Adult population
Tanaka 2014	20 (10/10)	20 (10/10)	A single dose (4g) of Fibrinogen (RiaSTAP®) given within 30 minutes of intervention decision.	One unit of apheresis platelets (median volume 230 ml) within 30 minutes of intervention decision.	Red blood cells, platelets, plasma, cryoprecipitate and fibrinogen if triggers met	i) Haemostatic condition in the surgical field post‐intervention; ii) Haemostatic blood product use over the first 24h; iii) Percentage of patients with thromboembolic events at 6‐8 weeks post‐operatively; iv) Mortality at 6‐8 weeks post‐operatively	Surgery (E): valve replacement or repair	Adult population
Innerhofer 2017	100 (52/48)	94 (50/44)	Fibrinogen concentrate (50 mg/kg of bodyweight) for abnormal ROTEM fibrin polymerisation (FibA10 < 9 mm) or four‐Factor PCC (20 IU/kg of bodyweight) for delayed initial thrombin formation (ExCT > 90s or prothrombin time index < 35%). FXIII concentrate (20 IU/kg of bodyweight) was administered with each second fibrinogen dose.	FFP (15 ml/kg bodyweight).	Tranexamic acid 20mg/kg, FXIII concentrate (20 IU/kg of bodyweight),	Occurrence of multiple organ failure during ICU stay as assessed by the daily SOFA score.	Trauma	Adult population

* 20 of the 47 patients had a 2nd dose of 20ml of PCC: in 3 because of continued bleeding and in 17 due to INR target not being met. 3 of the 47 patients received a 3rd dose of PCC due to an unmet INR level.

^ 3 of the 46 patients had a 2nd dose of 20 ml PCC: 1 due to a continued bleed and 2 did not reach the target INR.

d = days; (E) = elective surgery; FFP = fresh frozen plasma; h = hours; INR = International Normalised Ratio; mins = minutes; PCC = prothrombin complex concentrate; ROTEM^® = rotational thromboelastometry; TXA = tranexamic acid; (U) = urgent surgery/ treatment; wks = weeks.

Open in table viewer

Table 7. Factor XIII (FXIII): prophylactic trials with inactive comparator

Trial	Number randomised (N = intervention/ N = comparator)	Number included in analyses (N = intervention/ N = comparator)	Intervention	Comparator	Co‐interventions	Primary Outcomes	Clinical Condition	Population Age
Godje 2006	75 (25/25/ 25)	75 (25/25/25)*	Infusion of FactorXIII1) 1250 units or2) 2500 unitsimmediately after administration of protamine (protamine was administered during the scheduled coronary artery bypass graft surgery) This is the group that was included in this review.	NONE: standard surgical and post‐operative care	Protamine for heparin reversal FFP, red blood cells, platelets	Not stated	Surgery (E) : cardiac	Adult population
Karkouti 2013	479 (164/162/153)	409 (143/138/128)	A single intravenous dose of FactorXIII 1) 17.5 IU/kg or2) 35 IU/kgbefore the induction of anaesthesia. This is the group that was included in this review.	PLACEBO: single dose of an unspecified placebo before the induction of anaesthesia.	Protamine, antifibrinolytics during cardiopulmonary bypass, blood products	Percentage of patients avoiding any allogeneic transfusions within 7d or discharge home.	Surgery (E) : cardiac	Adult population
Korte 2009	25 (no further details stated)	22 (no further details stated)	FactorXIII (30 U/kg) infused 15 minutes after the beginning of surgery.	PLACEBO: albumin infusion.	Low molecular weight heparin, FFP, red blood cells, platelets	Clot firmness over 195 mins of surgery	Surgery (E): gastrointestinal cancer	Adult population
Levy 2009	43 (10/9/8/8)	43 (10/9/8/8)	A single infusion of FactorXIII at 4 doses:1) 11.9 units/kg;2) 25 units/kg;3) 35 units/kg;4) 50 units/kg. This is the group that was included in this review.	PLACEBO: no further details given.	Protamine for heparin reversal FFP, red blood cells, platelets Cryoprecipitate, Fibrinogen	Incidence and severity of adverse events from intervention administration to follow‐up at 5 to 7 wks	Surgery (E) : cardiac	Adult population
Rasche 1982	60 (29/31)	60 (29/31)	FactorXIII (1000 units) on day 1 of treatment and a subsequent daily dose of 500 units plus normal care protocols.	NONE: normal care products	Red blood cells, platelets, whole blood	Bleeding complications	Acute leukaemia	Adult population
Shirahata 1990	58 (30/28)	58 (30/28)	FactorXIII (dose of 70 to 100 units) administered slowly for at least 10 minutes intravenously within 6 hours of delivery.	NONE	Not stated	Not stated	Prematurity	Infant population

Open in table viewer

Table 8. Factor XIII (FXIII): therapeutic trials with inactive comparator

Trial

Number randomised (N = intervention/ N = comparator)

Number included in analyses (N = intervention/ N = comparator)

Intervention

Comparator

Co‐interventions

Primary Outcomes

Clinical Condition

Population Age

Bregenzer 1999

(17/11)

(11/9)

FactorXIII as an i.v. injection (at 3750 units on Day 0 and 1250 units on days 1 through 9) as an addition to basic steroid treatment.

PLACEBO: continuation of basic steroid treatment

TXA,

platelets or FFP post‐intervention

Time to cessation of macroscopically visible intestinal bleeding within 14d of starting treatment.

Ulceratice colitis

Adult population

Open in table viewer

Table 9. PCC: prophylactic trials with inactive comparator

Trial

Number randomised (N = intervention/ N = comparator)

Number included in analyses (N = intervention/ N = comparator)

Intervention

Comparator

Co‐interventions

Primary Outcomes

Clinical Condition

Population Age

Turner 1981

(39/39)

PCC (dose of 1ml/kg or 2ml/kg)

NONE: normal care protocols.

Vitamin K, heparin and clotting Factors for disseminated intravascular coagulation, red blood cells

i) Mortality (all‐cause)

ii) Incidence of intraventricular haemorrhage (post‐mortem)

Prematurity

Infant population

Setting

Three trials were multicentre, multinational trials (Karkouti 2013; Levy 2009; Rahe‐Meyer 2016) and six were multicentre trials, based in a single country: Germany (Bregenzer 1999), the UK (Collins 2017; Curry 2018), Spain (Sabate 2016), Sweden (Jeppsson 2016) and Denmark (Wikkelso 2015). The remaining 22 trials were single‐centre trials. Four trials were conducted in Iran (Najafi 2014; Sadeghi 2014; Pournajafian 2015; Soleimani 2016), three trials were conducted in Germany (Godje 2006; Rahe‐Meyer 2013; Rasche 1982), three trials were conducted in Switzerland (Haas 2015 (S); Haas 2015 (C); Korte 2009) and the Netherlands (Lance 2012; Bilecen 2017), whilst one trial was each conducted in Brazil (Galas 2014), Austria (Innerhofer 2017), Canada (Nascimento 2016), China (Cui 2010), Denmark (Fenger Erikson 2009), Great Britain (Turner 1981), Italy (Ranucci 2015), Japan (Shirahata 1990), Sweden (Karlsson 2011), and the USA (Tanaka 2014).

The clinical setting varied across the trials: 22 were in an elective surgical setting (Cui 2010; Fenger Erikson 2009; Galas 2014; Godje 2006; Haas 2015 (S) and (C); Karkouti 2013; Karlsson 2011; Korte 2009; Lance 2012; Levy 2009; Najafi 2014; Rahe‐Meyer 2013; Ranucci 2015; Sadeghi 2014; Tanaka 2014; Bilecen 2017; Jeppsson 2016; Pournajafian 2015; Rahe‐Meyer 2016; Sabate 2016; Soleimani 2016), five in an urgent medical setting (Wikkelso 2015; Collins 2017; Curry 2018; Innerhofer 2017; Nascimento 2016), and four in a non‐urgent medical setting (Bregenzer 1999; Rasche 1982; Shirahata 1990; Turner 1981). Definitions of elective and urgent have been taken from the trial manuscript.

Participants

Three trials were undertaken in infants (Haas 2015 (C); Shirahata 1990; Turner 1981), and three in paediatric populations (Cui 2010; Galas 2014; Haas 2015 (S)); the remaining 25 trials enrolled adult participants. We considered and analysed the infant trials separately for all outcomes other than allergic adverse events, where they have been grouped with the adult studies. We have included the paediatric population trials in analyses alongside the trials with adult populations, as we consider the populations in the three paediatric trials (Cui 2010, mean age 38.4 months; Haas 2015 (S), mean age of 12 years and Galas 2014, participants under 15 years of age) to be physiologically comparable to the populations in the trials recruiting adults.

Interventions

Prophylactic trials

Eighteen trials addressed prophylactic use of pro‐coagulant factors, 16 in comparison to inactive agents (Cui 2010; Fenger Erikson 2009; Godje 2006; Karkouti 2013; Karlsson 2011; Korte 2009; Levy 2009; Najafi 2014; Pournajafian 2015; Ranucci 2015; Rasche 1982; Sabate 2016; Sadeghi 2014; Shirahata 1990; Soleimani 2016; Turner 1981), and two in comparison to an active comparator (Haas 2015 (C); Haas 2015 (S)).

Inactive comparator trials

Of the 16 inactive comparator trials, seven recruited participants undergoing cardiac surgery (Cui 2010; Godje 2006; Karkouti 2013; Karlsson 2011; Levy 2009; Ranucci 2015; Sadeghi 2014). The other trials focused on elective cystectomy (Fenger Erikson 2009), gastrointestinal cancer (Korte 2009), acute leukaemia (Rasche 1982), spinal fusion (Pournajafian 2015), liver transplantation (Sabate 2016), trans‐urethral prostatic resection (Soleimani 2016), and elective hip replacement arthroplasty (Najafi 2014). Both neonatal population trials used inactive comparators and focused on babies at risk of intracranial haemorrhage (Shirahata 1990; Turner 1981).

Six of these prophylactic trials used factor XIII as the pro‐coagulant haemostatic intervention (Godje 2006; Karkouti 2013; Korte 2009; Levy 2009; Rasche 1982; Shirahata 1990), and nine used fibrinogen (Cui 2010; Fenger Erikson 2009; Karlsson 2011; Najafi 2014; Pournajafian 2015; Ranucci 2015; Sabate 2016; Sadeghi 2014; Soleimani 2016). Turner 1981 used a range of agents, including a prothrombin complex concentrate supplied by the Scottish National Blood Transfusion Service Protein Fractionation Centre: fibrinogen, factor VIII and platelets.

The inactive comparator agents were haemostatically inactive placebo products in 10 trials (Fenger Erikson 2009; Karkouti 2013; Korte 2009; Levy 2009; Najafi 2014; Pournajafian 2015; Ranucci 2015; Sabate 2016; Sadeghi 2014; Soleimani 2016), or nothing other than normal care (Cui 2010; Godje 2006; Karlsson 2011; Rasche 1982; Shirahata 1990; Turner 1981).

Active comparator trials

The participants in the two trials with an active comparator were neonates undergoing craniosynostosis or paediatric patients undergoing scoliosis surgery: both compared two doses of fibrinogen concentrate (Haas 2015 (C); Haas 2015 (S)).

Therapeutic trials

Thirteen trials addressed therapeutic use of pro‐coagulant factors: nine in comparison to inactive agents (Bilecen 2017; Bregenzer 1999; Collins 2017; Curry 2018; Jeppsson 2016; Nascimento 2016; Rahe‐Meyer 2013; Rahe‐Meyer 2016; Wikkelso 2015), and four compared to an active comparator (Galas 2014; Innerhofer 2017; Lance 2012; Tanaka 2014).

Inactive comparator trials

The inactive comparator trials focused on cardiac surgery (Bilecen 2017; Jeppsson 2016; Rahe‐Meyer 2013; Rahe‐Meyer 2016), ulcerative colitis (Bregenzer 1999), trauma (Curry 2018; Nascimento 2016), and postpartum haemorrhage (Collins 2017; Wikkelso 2015). The pro‐coagulant haemostatic interventions were fibrinogen in eight studies (Bilecen 2017; Collins 2017; Curry 2018; Jeppsson 2016; Nascimento 2016; Rahe‐Meyer 2013; Rahe‐Meyer 2016; Wikkelso 2015), and factor XIII (Bregenzer 1999). The comparators in all nine trials were haemostatically inactive placebo products.

Active comparator trials

The active comparator trials addressed cardiac surgery (Galas 2014; Tanaka 2014), trauma (Innerhofer 2017), and major elective surgery (Lance 2012). The intervention in all these trials was fibrinogen compared with FFP (Innerhofer 2017), cryoprecipitate (Galas 2014), platelets (Tanaka 2014), and as a partial replacement for FFP (Lance 2012).

Co‐interventions

Most trials documented the use of blood products (red blood cells and platelets) as well as transfusion of clotting products (FFP, prothrombin complex concentrate, cryoprecipitate) to treat bleeding or clotting derangement during the course of the trial.

Non‐blood product co‐interventions were reported by 26 trials. One trial reported routine administration of Vitamin K to all participants (Turner 1981), 11 trials reported administration of protamine (Bilecen 2017; Cui 2010; Godje 2006; Jeppsson 2016; Karkouti 2013; Karlsson 2011; Lance 2012; Levy 2009; Rahe‐Meyer 2013; Rahe‐Meyer 2016; Ranucci 2015), 14 trials reported the use of an antifibrinolytic agent (Bilecen 2017; Bregenzer 1999; Collins 2017; Curry 2018; Haas 2015 (C); Haas 2015 (S); Innerhofer 2017; Karkouti 2013; Nascimento 2016; Rahe‐Meyer 2013; Rahe‐Meyer 2016; Ranucci 2015; Sabate 2016; Wikkelso 2015) and two trials documented the use of post‐intervention low‐molecular weight heparin (Karlsson 2011; Korte 2009). Five trials did not report details of co‐interventions administered or outline additional transfusion policies (Galas 2014; Najafi 2014; Sadeghi 2014; Shirahata 1990; Soleimani 2016).

Primary outcomes of included trials

Twelve trials reported transfusion requirement, or the avoidance of transfusion, as a primary outcome (Collins 2017; Haas 2015 (C); Haas 2015 (S); Karkouti 2013; Najafi 2014; Rahe‐Meyer 2013; Rahe‐Meyer 2016; Ranucci 2015; Sabate 2016; Sadeghi 2014; Tanaka 2014; Wikkelso 2015). Four trials reported blood loss as the primary outcome (Bilecen 2017; Galas 2014; Jeppsson 2016; Soleimani 2016), and two trials each had primary outcomes of adverse events (Karlsson 2011; Levy 2009), or clot firmness following surgery (Fenger Erikson 2009; Korte 2009). Other primary outcomes were time to chest wall closure (Cui 2010), the incidence of multiorgan failure (Innerhofer 2017), mortality (Turner 1981), time to cessation of bleeding (Bregenzer 1999), clotting time and clot formation time (Lance 2012) and bleeding complications (Rasche 1982). Two trials were feasibility studies (Curry 2018; Nascimento 2016), and three studies did not report their primary outcome (Godje 2006; Pournajafian 2015; Shirahata 1990).

Funding sources

Seventeen trials received support from a commercial company for their study: CSL Behring (Bilecen 2017; Collins 2017; Curry 2018; Fenger Erikson 2009; Galas 2014; Haas 2015; Karlsson 2011; Lance 2012; Nascimento 2016; Rahe‐Meyer 2013; Rahe‐Meyer 2016; Ranucci 2015; Tanaka 2014), Novo Nordisk (Karkouti 2013; Korte 2009; Levy 2009), and Centeon Pharma GmbH (Bregenzer 1999). Eight of these trials also received support from an academic institution or charity (Bregenzer 1999; Collins 2017; Fenger Erikson 2009; Galas 2014; Karlsson 2011; Korte 2009; Nascimento 2016; Ranucci 2015).

Four studies were solely supported by grants from an academic institution or a charity (Jeppsson 2016; Sabate 2016; Sadeghi 2014; Wikkelso 2015). Eight trials did not report receipt of any funding (Cui 2010; Godje 2006; Innerhofer 2017; Najafi 2014; Rasche 1982; Shirahata 1990; Soleimani 2016; Turner 1981), whilst in one trial it was unclear whether or not funding had been received to support the trial (Pournajafian 2015).

Excluded studies

For the full list of excluded trials see Characteristics of excluded studies.

We excluded six trials as the participants were ineligible: they were on warfarin and received the pro‐coagulant haemostatic factor for anticoagulation reversal (Boulis 1999; Demeyere 2010; Goldstein 2015; Papakitsos 2014; Sarode 2013; van Aart 2006). We excluded one trial after translation from German and review of the full text for not being a randomised controlled trial (Scherer 1994). One study was excluded on review of the full text as it was a review of the literature (Maegele 2015). We excluded five studies as the study interventions did not meet the inclusion criteria for this review (Curry 2015; Garrigue 2017; Masoumi 2016; Miyata 2014; Qiu 2016). One study was excluded as it was terminated early for feasibility reasons after recruiting only a single participant to each study arm (TOP‐CLOT Trial).

We identified 22 trials as ongoing studies. Full details of these trials are available in the Characteristics of ongoing studies table.

Risk of bias in included studies

Please refer to Figure 2 for visual representations of the assessments of risk of bias across all trials and for each item in the included trials. See the 'Risk of bias' section in the Characteristics of included studies for more detailed information about the biases identified within the individual trials.

Figure 2

Risk of bias summary: review authors' judgements about each risk of bias item for each included trial. Thirty‐one trials are included in this review (Haas 2015 reports two trials).

Allocation

Random sequence generation

Prophylactic trials

We considered nine trials to be at low risk of bias, as randomisation was done centrally by a Voice response system (Karkouti 2013; Levy 2009), or was randomly generated (Korte 2009; Najafi 2014; Ranucci 2015; Sabate 2016; Sadeghi 2014; Soleimani 2016; Turner 1981). We judged nine trials to be at unclear risk of bias, as the method of randomisation was not adequately reported (Cui 2010; Fenger Erikson 2009; Godje 2006; Haas 2015 (C); Haas 2015 (S); Karlsson 2011; Pournajafian 2015; Rasche 1982; Shirahata 1990).

Therapeutic trials

We considered seven trials to be at low risk of bias, as randomisation was done centrally (Curry 2018; Rahe‐Meyer 2013; Wikkelso 2015), or was randomly generated (Bilecen 2017; Galas 2014; Innerhofer 2017; Nascimento 2016). We judged six trials to be at unclear risk of bias, as the method of randomisation was not adequately reported (Bregenzer 1999; Collins 2017; Jeppsson 2016; Lance 2012; Rahe‐Meyer 2016; Tanaka 2014).

Concealment of treatment allocation

Prophylactic trials

We considered five trials to be at low risk of bias for allocation concealment, as assignment was done centrally (Levy 2009; Sabate 2016) or was randomly generated (Haas 2015 (C); Haas 2015 (S); Korte 2009). We considered two trials to be at high risk of bias due to incorrect methods being used to conceal treatment allocation (Godje 2006; Ranucci 2015). We judged 11 trials to be at unclear risk of bias because no description of allocation concealment was provided (Cui 2010; Fenger Erikson 2009; Karkouti 2013; Karlsson 2011; Najafi 2014; Pournajafian 2015; Rasche 1982; Sadeghi 2014; Shirahata 1990; Soleimani 2016; Turner 1981).

Therapeutic trials

We considered seven trials to be at low risk of bias for allocation concealment, as assignment was done centrally (Rahe‐Meyer 2013) or was randomly generated (Bilecen 2017; Curry 2018; Galas 2014; Innerhofer 2017; Jeppsson 2016; Nascimento 2016). We considered three trials to be at high risk of bias, due to incorrect methods being used to conceal treatment allocation (Lance 2012; Tanaka 2014; Wikkelso 2015). We judged three trials to be at unclear risk of bias, because no description of allocation concealment was provided (Bregenzer 1999; Collins 2017; Rahe‐Meyer 2016).

Blinding

Blinding of participants and personnel (performance bias)

Prophylactic intervention

We considered 16 trials to be at low risk of participant performance bias as, i) participants were under anaesthesia when their intervention was administered (Cui 2010; Fenger Erikson 2009; Godje 2006; Haas 2015 (C); Haas 2015 (S) Karkouti 2013; Karlsson 2011; Korte 2009; Levy 2009; Pournajafian 2015; Ranucci 2015; Sabate 2016), ii) were too young to appreciate the nature of their intervention (Shirahata 1990; Turner 1981), iii) were unaware of the intervention content (Soleimani 2016), or iv) the plastic container vials used were covered and details of the contents not visible (Sadeghi 2014). We considered two trials, Najafi 2014 and Rasche 1982 to be at unclear risk of performance bias as insufficient information was provided to permit an alternative judgement. We did not consider any trials to be at high risk of participant performance bias.

We considered nine trials to be at low risk of personnel performance bias as: i) staff independent from the trial team prepared the interventions (Fenger Erikson 2009; Korte 2009; Sabate 2016; Soleimani 2016); ii) interventions were provided in visually identical containers (Levy 2009); iii) the plastic container vials used were covered and details of the contents were not visible (Sadeghi 2014); iv) clinical staff were blinded to treatment allocation (Karlsson 2011; Ranucci 2015); and v) outcome was not likely to be influenced by lack of blinding (Shirahata 1990). We considered six trials to be at high risk of personnel performance bias due to: i) the availability of the allocation concealment device in the participant's notes (Godje 2006); ii) the apparent non‐concealment of the notably visual difference between the intervention and control (Turner 1981); iii) personnel being unblinded to treatment allocation (Haas 2015 (C); Haas 2015 (S)); and iv) a lack of information on blinding where knowledge of treatment allocation could introduce performance bias (Cui 2010; Karkouti 2013). We considered three trials to be at unclear risk of personnel performance bias, as insufficient information was provided to permit an alternative judgement (Najafi 2014; Pournajafian 2015; Rasche 1982).

Therapeutic trials

We considered all 13 trials to be at low risk of participant performance bias because: i) intervention and placebo were identical in composition (Bregenzer 1999; Collins 2017; Curry 2018; Nascimento 2016; Wikkelso 2015); ii) participants were under anaesthesia when their intervention was administered with the trial personnel and clinicians blinded to treatment allocation (Bilecen 2017; Galas 2014; Jeppsson 2016; Lance 2012; Rahe‐Meyer 2013; Rahe‐Meyer 2016; Tanaka 2014); and iii) participants had a low level of consciousness or were unconscious at the time of the study intervention, although the clinicians administering the intervention were not blinded to its nature (Innerhofer 2017).

We considered nine trials to be at low risk of personnel performance bias because the placebo was identical in composition to the intervention (Bilecen 2017; Bregenzer 1999; Collins 2017; Curry 2018; Jeppsson 2016; Nascimento 2016; Rahe‐Meyer 2013; Rahe‐Meyer 2016; Wikkelso 2015). We considered two trials to be at high risk of personnel performance bias as these were open‐label trial designs (Innerhofer 2017; Tanaka 2014), and we considered two trials to be at an unclear risk of performance bias, as insufficient information was provided to permit an alternative judgement (Galas 2014; Lance 2012).

Blinding of outcome assessors (detection bias)

Prophylactic intervention

We considered six trials to be at low risk of detection bias, as outcome assessments were: i) made by trial personnel blinded to treatment allocation (Cui 2010; Fenger Erikson 2009; Karkouti 2013; Karlsson 2011); or ii) knowledge of treatment allocation would not affect assessment of objective outcomes (Rasche 1982; Shirahata 1990). We considered two trials to be at high risk of detection bias due to lack of appropriate allocation concealment (Godje 2006; Turner 1981), as well as notable functional and visual differences between the intervention and control in Turner 1981. We considered 10 trials to be at unclear risk of detection bias, as insufficient information was provided to permit an alternative judgement (Haas 2015 (C); Haas 2015 (S); Korte 2009; Levy 2009; Najafi 2014; Pournajafian 2015; Ranucci 2015; Sabate 2016; Sadeghi 2014; Soleimani 2016).

Therapeutic trials

We considered six trials to be at low risk of detection bias as outcomes were: i) measured by participants blinded to their treatment allocation and by an independent reference laboratory (Bregenzer 1999); ii) trial personnel were blinded to treatment allocation (Bilecen 2017; Galas 2014; Rahe‐Meyer 2013; Wikkelso 2015); and iii) knowledge of allocation would not influence assessment of the trial's objective outcomes (Lance 2012). We considered two trials to be at high risk of detection bias as these were open‐label trial designs (Innerhofer 2017; Tanaka 2014). We considered five studies to be at unclear risk of detection bias, as insufficient information was provided to permit a judgement of high or low risk of detection bias (Collins 2017; Curry 2018; Jeppsson 2016; Nascimento 2016; Rahe‐Meyer 2016).

Incomplete outcome data

Prophylactic intervention

We considered 11 trials to be at low risk of attrition bias, with all participants accounted for within the study flow diagram and no withdrawals (Fenger Erikson 2009; Godje 2006; Haas 2015 (C); Haas 2015 (S); Karkouti 2013; Najafi 2014; Ranucci 2015; Sadeghi 2014; Turner 1981), or where participants were lost to follow‐up this was evenly spread across trial arms (Sabate 2016; Soleimani 2016). We considered two trials to be at high risk of attrition bias because no details were reported about the intervention groups of participants excluded from analysis (Levy 2009) and four participants were excluded at unclear phases of the study, but following recruitment and may have been excluded following randomisation and treatment administration (Pournajafian 2015). We considered five trials to be at unclear risk of attrition bias because: i) the issue of participant dropout and post‐randomisation exclusion was not addressed (Cui 2010; Karlsson 2011; Rasche 1982; Shirahata 1990); and ii) the number of participants randomised to each treatment arm or analysed by outcome was not reported (Korte 2009).

Therapeutic trials

We considered 11 trials to be at low risk of attrition bias, as all participants were accounted for within the study flow diagram and there were no withdrawals (Collins 2017; Curry 2018; Galas 2014; Innerhofer 2017; Jeppsson 2016; Lance 2012; Nascimento 2016; Rahe‐Meyer 2013; Rahe‐Meyer 2016; Tanaka 2014; Wikkelso 2015). We considered one trial to be at high risk of attrition bias, because a third of all participants were not available for the intention‐to‐treat analyses (Bregenzer 1999). We considered one trial to be at unclear risk of attrition bias, because we determined that the missing data forms from participants in both study arms were unlikely to have an impact on outcome; as the data were balanced across study groups, this presented an unclear risk of bias (Bilecen 2017).

Selective reporting

Prophylactic trials

We considered 14 trials to be at low risk of reporting bias, because all the outcomes were prespecified by the authors and provided in the results (Cui 2010; Fenger Erikson 2009; Godje 2006; Haas 2015 (C); Haas 2015 (S); Karkouti 2013; Karlsson 2011; Korte 2009; Ranucci 2015; Rasche 1982; Sabate 2016; Sadeghi 2014; Soleimani 2016; Turner 1981). We considered two trials to be at high risk of reporting bias, as the trials did not report data for all predefined outcomes (Levy 2009; Najafi 2014). We considered two trials to be at unclear risk of reporting bias, as outcomes were not predefined (Pournajafian 2015; Shirahata 1990).

Therapeutic trials

We considered nine trials to be at low risk of reporting bias, because all the outcomes were prespecified by the authors and covered in the results (Bilecen 2017; Collins 2017; Galas 2014; Jeppsson 2016; Nascimento 2016; Rahe‐Meyer 2013; Rahe‐Meyer 2016; Tanaka 2014; Wikkelso 2015).

We considered two trials to be at high risk of reporting bias (Bregenzer 1999; Innerhofer 2017). In Bregenzer 1999, due to the early discontinuation of the study, data were not reported for a number of prespecified outcomes, including haemoglobin and haematocrit levels. In Innerhofer 2017, the published manuscript did not report findings from all the outcomes they purported to measure. We considered two trials to be at unclear risk of reporting bias, as there was no trial protocol and no clarity about the outcomes measured in the trial (Lance 2012). In Curry 2018, the study protocol outlined the use of ROTEM data, but none were reported.

Other potential sources of bias

Prophylactic trials

We determined four studies to be at low risk of other potential sources of bias in study design; Fenger Erikson 2009 appeared free of all concerns, while Haas 2015 (C), Haas 2015 (S) and Karlsson 2011 declared industrial funding and support, but reported no conflicts of interest or funder control over data or publication. We considered one trial (Karkouti 2013) to be at high risk of bias due to variations in transfusion practice, despite declared protocols and the fact that the company producing the product and sponsoring the trial oversaw the trial operations, audited all data, and conducted the statistical analyses. There is also unclear ownership of trial data, although the publication states full author access.

We determined that 13 trials were at an unclear risk of other bias. In seven of these trials, the reasoning was: i) participants enrolled in Cui 2010 appear to have been randomised prior to application of exclusion criteria with more participants excluded from the intervention arm of the study, although the allocation and exclusion numbers were not stated; ii) Godje 2006 did not declare sources of funding or a statement of author interests; iii) Korte 2009 declared industry funding with an industrial employee on the byline, but did not provide any indication of roles or independence of publication or data ownership; iv) Levy 2009 declared industrial funding of the study and three authors, including the senior author, were industrial employees. However there was no statement of data ownership, study independence or industrial control of data, manuscript content or publication; v) Ranucci 2015 declared full author access to data and decision to publish, independent of industrial funding, but ownership of the data was unclear; vi) Sabate 2016 provided a power analysis identifying that 66 participants were required in each study group, hence insufficient participants (99) were recruited for the study's primary outcome conclusions; and vii) the source of funding for Turner 1981 was unclear. In the other six trials there was insufficient information from which to determine the risk of any other forms of bias (Najafi 2014; Pournajafian 2015; Rasche 1982; Sadeghi 2014; Shirahata 1990, Soleimani 2016).

Therapeutic trials

We deemed six therapeutic trials to be at low risk of other sources of bias. Five trials (Bilecen 2017; Galas 2014; Innerhofer 2017; Jeppsson 2016; Lance 2012) appeared free of all concerns, whilst the sixth (Collins 2017) declared industrial support for the study but provided detailed information about involvement and low industrial influence.

We considered two trials to be at high risk of bias, due to a declaration of funding by CS Behring and having industrial employees as authors without providing any information about data ownership or control, study, manuscript and publication independence (Rahe‐Meyer 2013; Rahe‐Meyer 2016). Five trials did not provide sufficient information for us to determine risk of other potential sources of bias (Bregenzer 1999; Curry 2018; Nascimento 2016; Tanaka 2014; Wikkelso 2015).

Effects of interventions

See: Summary of findings for the main comparison FIBRINOGEN: Therapeutic trials: intervention compared to inactive Control for the prevention and treatment of bleeding in participants without haemophilia; Summary of findings 2 FIBRINOGEN: Therapeutic trials: intervention compared to haemostatically active control for the prevention and treatment of bleeding in participants without haemophilia; Summary of findings 3 FIBRINOGEN: Prophylactic trials: intervention compared to inactive control for the prevention and treatment of bleeding in participants without haemophilia; Summary of findings 4 FIBRINOGEN: Prophylactic trials: intervention compared to haemostatically active control for the prevention and treatment of bleeding in participants without haemophilia; Summary of findings 5 FXIII: Prophylactic trials: intervention compared to inactive control for the prevention and treatment of bleeding in participants without haemophilia

We present the outcome data in the following way.

by outcome.
by type of intervention.
by whether the intervention was used in a prophylactic or therapeutic setting.
by whether the intervention was being compared with a haemostatically inactive comparator (for example, placebo or saline) or a haemostatically active comparator (whether a different dose of the same intervention or a different haemostatically active intervention).

However, the paucity of quantifiable data that can be combined in a meta‐analysis is such that we cannot provide quantitative analysis for most comparisons.

We analysed the three trials enrolling neonates separately from the other included trials (as previously described) (Haas 2015 (C); Shirahata 1990; Turner 1981). Due to the small volumes of the placebo agents and their inactive nature in two trials (Shirahata 1990; Turner 1981), we have compared all trials in this group as if their comparators are equivalent.

Three trials assessed the prophylactic use of factor XIII against an inactive comparator with multiple factor XIII dosing arms (Godje 2006; Karkouti 2013; Levy 2009). While we appreciate it is not possible to generate an exact comparison across these trials, to facilitate some meaningful analysis we have included the data for the 2500 unit (Godje 2006) and 35 unit/kg arms (Karkouti 2013; Levy 2009), as these should correlate for a 'typical' 70 kg person. Each trial included only a single control arm and we have included them in full.

In Fenger Erikson 2009, the average participant weight was 77 kg, and we have used this value to calculate total intervention doses given.

In Korte 2009, the breakdown of the 22 randomised participants was not reported by intervention group, but based on graphical data in the report we have been able to identify that 11 participants were included in each study arm.

Primary outcome: mortality (all‐cause)

Fibrinogen concentrate: prophylactic trials with inactive comparator

(Nine trials: Cui 2010; Fenger Erikson 2009; Karlsson 2011; Najafi 2014; Pournajafian 2015; Ranucci 2015; Sabate 2016; Sadeghi 2014; Soleimani 2016)

Four trials comparing fibrinogen concentrate versus inactive comparators with 248 participants reported data on all‐cause mortality (Fenger Erikson 2009; Karlsson 2011; Ranucci 2015; Sabate 2016), but we did not pool data due to differences in the timing of the outcome measurement.

There was no evidence of a difference in the number of deaths between the fibrinogen and the inactive comparator arms: risk ratio (RR) 0.33, 95% confidence interval (CI) 0.04 to 3.11; in 116 participants up to 30 days post‐complex cardiac surgery (Ranucci 2015) and RR 0.31, 95% CI 0.03 to 2.83; in 92 participants prior to hospital discharge after liver transplantation (Sabate 2016). Fenger Erikson 2009 and Karlsson 2011 reported no events at up to four days after radical cystectomy amongst 20 participants and up to seven days after coronary artery bypass graft surgery amongst 20 participants, respectively (Analysis 1.1).

Fibrinogen concentrate: prophylactic trials with active comparator

(Two participant groups reported in 1 trial: Haas 2015 (C); Haas 2015 (S))

No mortality was reported by either Haas 2015 (C) or Haas 2015 (S), with 49 participants in total when comparing two different thresholds for administering prophylactic fibrinogen concentrate in complex surgery (Analysis 2.1).

Fibrinogen concentrate: therapeutic trials with inactive comparator

(Eight trials: Bilecen 2017; Collins 2017; Curry 2018; Jeppsson 2016; Nascimento 2016; Rahe‐Meyer 2013; Rahe‐Meyer 2016; Wikkelso 2015)

All‐cause mortality data from 724 participants was reported by seven trials comparing fibrinogen concentrate with an inactive agent in the management of active bleeding (Bilecen 2017; Collins 2017; Curry 2018; Nascimento 2016; Rahe‐Meyer 2013; Rahe‐Meyer 2016; Wikkelso 2015). Differences in the timing of the outcome measure and clinical setting of the trial prevented us from pooling data across the seven trials (Analysis 3.1).

Bilecen 2017 did not demonstrate any evidence of a difference in in‐hospital mortality within 30 days of cardiac surgery between the fibrinogen and inactive comparator intervention arms: RR 5.00, 95% CI 0.25 to 102.00; 120 participants; low‐quality evidence (summary of findings Table for the main comparison).

Pooled data at 46 days following aortic valve surgery (Rahe‐Meyer 2013) and complex cardiovascular surgery (Rahe‐Meyer 2016) did not demonstrate any evidence of a difference in the number of deaths between the fibrinogen and inactive comparator intervention arms: RR 0.23, 95% CI 0.05 to 1.01; 213 participants; high‐quality evidence (summary of findings Table for the main comparison).

Two trials investigating the role of fibrinogen concentrate in the management of trauma haemorrhage (Curry 2018; Nascimento 2016) did not demonstrate any evidence of a difference in the number of deaths between the fibrinogen and inactive comparator intervention arms at 28 days following admission: RR 1.46, 95% CI 0.71 to 2.99; 97 participants; moderate‐quality evidence (summary of findings Table for the main comparison).

In the management of postpartum haemorrhage no deaths were reported or recorded in two trials six weeks postnatally (Collins 2017; 50 participants and Wikkelso 2015; 244 participants).

Fibrinogen concentrate: therapeutic trials with active comparator

(Four trials: Galas 2014; Innerhofer 2017; Lance 2012; Tanaka 2014)

Four studies comparing fibrinogen concentrate to haemostatically active comparators reported on all‐cause mortality in 220 participants at different time points (Galas 2014; Innerhofer 2017; Lance 2012; Tanaka 2014). No studies reported comparable time points, hence we did not pool data in a meta‐analysis across the trials (Analysis 4.1).

Following major elective surgery, there was no evidence of a difference in the number of deaths between fibrinogen and FFP at post‐operative day 30 (Lance 2012: RR 0.95, 95% CI 0.06 to 14.30; 43 participants, low‐quality evidence (summary of findings Table 2)) or 30 days after major trauma (Innerhofer 2017: RR 2.20, 95% CI 0.45 to 10.78; 94 participants, very low‐quality evidence (summary of findings Table 2)).

Following cardiac surgery there were no recorded deaths in either the fibrinogen or cryoprecipitate arms at post‐operative day seven or hospital discharge (Galas 2014: 63 participants) or when fibrinogen concentrate was compared to platelet infusion at 28 days post‐operatively (Tanaka 2014: 20 participants).

Factor XIII (FXIII): prophylactic trials with inactive comparator

(Six trials: Godje 2006; Karkouti 2013; Korte 2009; Levy 2009; Rasche 1982; Shirahata 1990)

All‐cause mortality was reported for 414 participants in five trials comparing prophylactic factor XIII with a haemostatically inactive control in adult populations (Godje 2006; Karkouti 2013; Korte 2009; Levy 2009; Rasche 1982) (Analysis 5.1).

Following coronary artery bypass graft surgery there was no evidence of a difference in the number of deaths between the factor XIII and placebo intervention arms at five to seven weeks post‐operatively in Karkouti 2013: RR 2.78, 95% CI 0.11 to 67.73; 266 participants. Levy 2009 reported no events in 16 participants who had undergone coronary artery bypass graft surgery. Godje 2006 recorded no deaths prior to hospital discharge in 50 participants following cardiac surgery.

In non‐cardiac scenarios there was no evidence of a difference in all‐cause mortality between factor XIII and haemostatically inactive control groups at an unspecified time point in acute leukaemia patients in Rasche 1982: RR 1.07, 95% CI 0.49 to 2.32; 60 participants; or at a median follow‐up of 340 days from elective gastrointestinal surgery in Korte 2009: RR 1.00, 0.26 to 3.91; 22 participants.

Factor XIII (FXIII): therapeutic trials with inactive comparator

(one trial: Bregenzer 1999)

Bregenzer 1999 did not report on all‐cause mortality.

Prothrombin complex concentrate (PCC): prophylactic trials with inactive comparator

(one trial: Turner 1981)

In Turner 1981 there was no evidence of a difference in the number of deaths amongst pre‐term neonates (time point not stated) when PCC was compared to no treatment: RR 1.05, 95% CI 0.71 to 1.53; 78 participants (Analysis 7.1).

Primary outcome: arterial and venous thromboembolic events (VTE)

Fibrinogen concentrate: prophylactic trials with inactive comparator

(Nine trials: Cui 2010; Fenger Erikson 2009; Karlsson 2011; Najafi 2014; Pournajafian 2015; Ranucci 2015; Sabate 2016; Sadeghi 2014; Soleimani 2016)

Seven trials reported the incidence of arterial events in 398 participants where fibrinogen concentrate was given prophylactically in comparison to an inactive control (Fenger Erikson 2009; Karlsson 2011; Najafi 2014; Ranucci 2015; Sabate 2016; Sadeghi 2014; Soleimani 2016) (Analysis 1.2; Analysis 1.3). We did not combine data across all studies due to differences in the time point of measurement and the clinical settings of the trials.

Arterial events

In participants undergoing cardiac surgery, at four days post‐operatively Sadeghi 2014 reported no arterial events in either study arm while Karlsson 2011 reported a single event in each arm (60 and 20 participants, respectively). At 30 days following cardiac surgery, there was no evidence of a difference in the number of arterial events in Ranucci 2015: RR 0.33, 95% CI 0.01 to 8.02; 116 participants; low‐quality evidence (summary of findings Table 3).

In studies that assessed fibrinogen concentrate in elective surgery, no arterial events were reported up to three days following radical cystectomy in Fenger Erikson 2009 (20 participants), up to one week after trans‐urethral prostatic resection in Soleimani 2016 (60 participants) or at 30 days following hip replacement surgery in Najafi 2014 (30 participants). In Sabate 2016, there was no evidence of a difference in arterial event incidence between the intervention arms 30 days after liver transplantation: RR 0.31, 95% 0.03 to 2.83; 92 participants, low‐quality evidence (summary of findings Table 3).

Venous events

Following cardiac surgery, there was no evidence of a difference in the incidence of venous thromboembolic complications at four days between the intervention arms in Karlsson 2011: RR 3.00, 95% CI 0.14 to 65.90; 20 participants, moderate‐quality evidence (summary of findings Table 3).

No venous events were reported at four days following cardiac surgery in 60 participants in Sadeghi 2014 or amongst 116 participants at 30 days following cardiac surgery in Ranucci 2015. Following elective surgery Fenger Erikson 2009 (20 participants), Soleimani 2016 (60 participants) and Najafi 2014 (30 participants) did not report any venous events at three, seven and 30 days, respectively.

In Sabate 2016 there was no evidence of a difference in the incidence of venous thromboembolic complications between the intervention arms 30 days following liver transplantation: RR 0.31, 95% CI 0.01 to 7.32; 92 participants, low‐quality evidence (summary of findings Table 3).

Fibrinogen concentrate: prophylactic trials with active comparator

(Two trials reported in 1 study: Haas 2015 (C); Haas 2015 (S))

Haas 2015 (C) and Haas 2015 (S) reported no arterial or venous thromboembolic events in either fibrinogen concentrate supplementation arms (total 49 participants) during the entire participant hospital stay.

Fibrinogen concentrate: therapeutic trials with inactive comparator

(Eight trials: Bilecen 2017; Collins 2017; Curry 2018; Jeppsson 2016; Nascimento 2016; Rahe‐Meyer 2013; Rahe‐Meyer 2016; Wikkelso 2015)

Six trials reported both arterial and venous thromboembolic events at a single time point each in a total of 562 participants (Bilecen 2017; Collins 2017; Curry 2018; Jeppsson 2016; Rahe‐Meyer 2013; Wikkelso 2015). Additionally, adverse event data from Nascimento 2016 (45 participants) could be used to calculate the per capita event rate for arterial events but not venous events (Analysis 3.2; Analysis 3.3).

Arterial events

Neither Collins 2017 nor Wikkelso 2015 found any arterial events at six weeks following postpartum haemorrhage in 50 and 244 participants, respectively. There were no reported arterial events in Nascimento 2016 at 28 days following admission for trauma (45 participants). Curry 2018 found no evidence of a difference between the fibrinogen and control arms in the arterial events rate at 28 days following admission for trauma: RR 0.47, 95% CI 0.05 to 4.82; 39 participants.

In the cardiac surgery studies, there was no evidence of any differences in the number of arterial events at post‐operative day 30 (Bilecen 2017; RR 2.33, 95% CI 0.63 to 8.60; 120 participants), day 45 (Rahe‐Meyer 2013: RR 1.10, 95% CI 0.07 to 16.85; 61 participants) and during hospital stay (Jeppsson 2016: RR 0.33, 95% CI 0.01 to 7.80; 48 participants).

Venous events

There were no recorded venous thromboembolic events following cardiac surgery in Bilecen 2017 (120 participants) or Jeppsson 2016 (48 participants), or at six weeks following postpartum haemorrhage in Wikkelso 2015 (244 participants). In the three trials recording venous thromboembolic events there was no evidence of a difference in the event rate between the fibrinogen and inactive comparator intervention arms: following cardiac surgery in Rahe‐Meyer 2013: RR 0.37, 95% CI 0.02 to 8.66; 61 participants; at six weeks following postpartum haemorrhage in Collins 2017: RR 0.92, 95% CI 0.06 to 13.95), 50 participants; and 28 days after an admission with major trauma in Curry 2018: RR 4.76, 95% CI 0.24 to 93.19; 39 participants.

Fibrinogen concentrate: therapeutic trials with active comparator

(Four trials: Galas 2014; Innerhofer 2017; Lance 2012; Tanaka 2014)

Three studies reported on both arterial and venous complications in 126 participants receiving therapeutic fibrinogen concentrate or active comparator (Galas 2014; Lance 2012; Tanaka 2014). Additionally Innerhofer 2017 reported on the incidence of venous thromboembolic complications in a further 94 participants. We did not pool data collectively due to differences in the timing of the outcome measurement (Analysis 4.2; Analysis 4.3), the active comparator and the clinical setting of the trials.

Arterial events

There was no evidence of a difference in arterial event incidence between the fibrinogen and active comparator arms in Galas 2014 seven days after cardiac surgery: RR 0.44, 95% CI 0.09 to 2.10; 63 participants; or in Tanaka 2014 28 days after cardiac surgery: RR 0.33, 95% CI 0.02 to 7.32; 20 participants; or in Lance 2012 30 days after major elective surgery: RR 2.87, 95% CI 0.12 to 66.75; 43 participants.

Venous events

No venous events were reported in either cardiac surgery trial (Galas 2014; Tanaka 2014).

There was no evidence of a difference in venous thromboembolic event incidence between the fibrinogen and active comparator arm in Innerhofer 2017 30 days after trauma admission: RR 0.63, 95% CI 0.21 to 1.87; 94 participants; or following major surgery in Lance 2012: RR 3.00, 95% CI 0.12 to 77.83; 43 participants.

Factor XIII (FXIII): prophylactic trials with inactive comparator

(six trials: Godje 2006; Karkouti 2013; Korte 2009; Levy 2009; Rasche 1982; Shirahata 1990)

Four trials of factor XIII prophylaxis reported on thromboembolic complications in 338 participants in cardiac surgery (Godje 2006; Karkouti 2013; Levy 2009) and surgery for gastrointestinal cancer (Korte 2009) (Analysis 5.2; Analysis 5.3)

Arterial events

No arterial events were reported at up to seven weeks following cardiac surgery in Levy 2009 (16 participants).

There was no evidence of a difference in arterial thromboembolic events in Karkouti 2013 up to seven weeks following cardiac surgery: RR 0.67, 95% CI 0.28 to 1.62;, 266 participants; low‐quality evidence; or in Godje 2006 at an unspecified post‐operative time point: RR 0.20, 95% CI 0.01 to 3.97; 50 participants; very low‐quality evidence; or Korte 2009 following gastrointestinal surgery: RR 3.00, 95% CI 0.14 to 66.53; 22 participants; low‐quality evidence.

Venous events

No venous events were reported at up to seven weeks following cardiac surgery in Levy 2009 (16 participants) or Godje 2006 (50 participants).

There was no evidence of a difference in venous thromboembolic events in Karkouti 2013 up to seven weeks following cardiac surgery: RR 0.93, 95% CI 0.06 to 14.67; 266 participants; low‐quality evidence; or Korte 2009 following gastrointestinal surgery: RR 3.00, 95% CI 0.14 to 66.53; 22 participants; low‐quality evidence.

Factor XIII (FXIII): therapeutic trials with inactive comparator

(one trial: Bregenzer 1999)

Bregenzer 1999 did not report the incidence of arterial or venous thromboembolic complications.

Prothrombin complex concentrate (PCC): prophylactic trials with inactive comparator

(One trial; Turner 1981)

Turner 1981 did not report the incidence of arterial or venous thromboembolic complications.

Secondary outcome: mortality due to bleeding

Fibrinogen concentrate: prophylactic trials with inactive comparator

(Nine trials: Cui 2010; Fenger Erikson 2009; Karlsson 2011; Najafi 2014; Pournajafian 2015; Ranucci 2015; Sabate 2016; Sadeghi 2014; Soleimani 2016)

Fenger Erikson 2009, Karlsson 2011; Ranucci 2015; and Sabate 2016 reported on this outcome in 248 participants. There was only a single death due to bleeding reported in the fibrinogen arm of Sabate 2016: RR 2.76, 95% CI 0.12 to 65.92; 92 participants (Analysis 1.4).

Fibrinogen concentrate: prophylactic trials with active comparator

(Two participant groups reported in one trial: Haas 2015 (C); Haas 2015 (S))

No deaths due to bleeding were reported in Haas 2015 (C) or Haas 2015 (S) from 49 participants in either fibrinogen supplementation arm (Analysis 2.4).

Fibrinogen concentrate: therapeutic trials with inactive comparator

(Eight trials: Bilecen 2017; Collins 2017; Curry 2018; Jeppsson 2016; Nascimento 2016; Rahe‐Meyer 2013; Rahe‐Meyer 2016; Wikkelso 2015)

Five studies reported data on death due to haemorrhage with a total of 539 participants. Rahe‐Meyer 2013 reported no deaths due to bleeding occurred following cardiac surgery in 152 participants up to post‐operative day 46 (Analysis 3.4).

Following postpartum haemorrhage, no deaths due to bleeding were reported at six weeks postnatally in either Collins 2017 or Wikkelso 2015 in 50 and 244 participants, respectively.

Mortality due to bleeding following administration of therapeutic fibrinogen at 28 days following trauma admission was reported and pooled for Curry 2018 and Nascimento 2016: there was no evidence of any difference in the number of deaths between the intervention arms: RR 2.45, 95% CI 0.38 to 15.76; 93 participants.

Fibrinogen concentrate: therapeutic trials with active comparator

(Four trials: Galas 2014; Innerhofer 2017; Lance 2012; Tanaka 2014).

Four trials with 220 participants reported no incidents of mortality due to bleeding to post‐cardiac surgery day seven (Galas 2014: 63 participants) and day 28 (Tanaka 2014: 20 participants; Lance 2012; 43 participants), or at 30 days following trauma (Innerhofer 2017: 94 participants) (Analysis 4.4).

Factor XIII (FXIII): prophylactic trials with inactive comparator

(Six trials: Godje 2006; Karkouti 2013; Korte 2009; Levy 2009; Rasche 1982; Shirahata 1990)

Four trials with 392 participants reported data for this outcome (Godje 2006; Karkouti 2013; Levy 2009; Rasche 1982) (Analysis 5.4).

Rasche 1982 reported mortality due to bleeding at a median of 32 days following the study intervention: there was no evidence of a difference in mortality due to bleeding in people with acute leukaemia given factor XIII infusion compared to no intervention: RR 0.80, 95% CI 0.20 to 3.28; 60 participants.

Godje 2006; Karkouti 2013; Levy 2009 reported on this outcome following cardiac surgery and reported that no deaths due to bleeding occurred (total 332 participants).

Factor XIII (FXIII): therapeutic trials with inactive comparator

(One trial:Bregenzer 1999)

Bregenzer 1999 did not report on this outcome.

Prothrombin complex concentrate (PCC): prophylactic trials with inactive comparator

(One trial: Turner 1981).

Turner 1981 did not report on this outcome.

Secondary outcome: red blood cell (RBC) transfusion requirement

See Table 1 for individual study data for this outcome, which was reported in a number of ways (means, median, total) across the included studies.

Fibrinogen concentrate: prophylactic trials with inactive comparator

(Nine trials:Cui 2010; Fenger Erikson 2009; Karlsson 2011; Najafi 2014; Pournajafian 2015; Ranucci 2015; Sabate 2016; Sadeghi 2014; Soleimani 2016)

Mean red blood cell volume required per participant

Karlsson 2011, Najafi 2014 and Sadeghi 2014 (110 participants in total) reported these data as mean red blood cell volume required per participant, but pooling of data was not possible due to the lack of common time points. Following cardiac surgery, fibrinogen concentrate reduced the red blood cell transfusion requirement at 12 hours post‐operatively (Karlsson 2011; mean difference (MD) −150.00, 95% CI −205.72 to −94.28 mL; 20 participants; although this effect was not reproduced intra‐operatively (MD −90.00, 95% CI −281.13 to 101.13 mL; 60 participants). There was no evidence of a difference in red blood cell requirement between the intervention arms at 24 hours post‐cardiac surgery in Sadeghi 2014 (MD −150.00, 95% CI −401.53 to 101.53 mL; 60 participants) or in the perioperative period following hip replacement surgery (Najafi 2014; MD −78.00, 95% CI −317.29 to 161.29 mL; 30 participants) (Analysis 1.5).

Receipt of a red blood cell transfusion

Nine studies (with 470 participants) reported the number of participants receiving a red cell transfusion (Cui 2010; Fenger Erikson 2009; Karlsson 2011; Najafi 2014; Pournajafian 2015; Ranucci 2015; Sabate 2016; Sadeghi 2014; Soleimani 2016).

Following cardiac surgery three studies provided a common time point at 24 hours post‐operatively and we pooled the data (Cui 2010; Ranucci 2015; Sadeghi 2014). The data showed a reduction in the number of participants receiving a RBC transfusion with the use of prophylactic fibrinogen in cardiac surgery with moderate heterogeneity: RR 0.53, 95% CI 0.37 to 0.76; P < 0.001; 207 participants; I² = 31% (Analysis 1.6). This association was also present when we used a random‐effects model. We noted no change to the effect estimate, but an increase in heterogeneity (analysis not shown) when we restricted the analysis to:

trials including adult participants only: RR 0.53, 95% CI 0.36 to 0.78, I² = 65%; 176 participants) (Ranucci 2015; Sadeghi 2014);
trials where the number of participants who were not included in the analysis was unclear: RR 0.53, 95% CI 0.36 to 0.78, I² = 65%; 176 participants) (Ranucci 2015; Sadeghi 2014).

We noted a change in the effect estimate in that there was no difference in blood loss between the intervention arms and a reduction in heterogeneity when
we restricted the analysis to to trials with a low or uncertain risk of bias: RR 0.67, 95% CI 0.44 to 1.01, I² = 0%; 91 participants) (Cui 2010; Sadeghi 2014).

We also found evidence for a reduction in the number of participants receiving a transfusion between the fibrinogen and inactive comparator arm at 48 hours post‐operatively in Ranucci 2015: RR 0.59, 95% CI 0.38 to 0.92; 116 participants. Conversely, Karlsson 2011 did not find any evidence for a reduction in the number of participants requiring a red blood cell transfusion at 12 hours following coronary artery bypass grafting: RR 0.33, 95% CI 0.04 to 2.69; 20 participants.

We pooled data from the elective, non‐cardiac surgery trials (Fenger Erikson 2009; Soleimani 2016), and found evidence of a reduction in the number of participants requiring a red blood cell transfusion following prophylactic fibrinogen at 48 hours post‐operatively: RR 0.23, 95% CI 0.07 to 0.68; P = 0.008; 80 participants; I² = 0%; Analysis 1.6). In the trial of participants undergoing spinal fusion surgery (Pournajafian 2015) and liver transplantation (Sabate 2016) there was no evidence of a difference in the number of participants requiring a RBC transfusion between the intervention arms at 24 hours following surgery: RR 1.01, 95% CI 0.76 to 1.33; 92 participants; and RR 0.07, 95% CI 0.00 to 1.22; 41 participants, respectively).

Four studies reported data at a range of different time points, as medians with IQRs (Cui 2010; Fenger Erikson 2009; Ranucci 2015; Sabate 2016). See Table 1 for these values.

Fibrinogen concentrate: prophylactic trials with active comparator

(Two participant groups reported in one trial: Haas 2015 (C); Haas 2015 (S))

Haas 2015 (C) and Haas 2015 (S) reported the mean and median quantity of red cells transfused (in 49 participants). This trial compared receipt of fibrinogen concentrate at two predefined intraoperative fibrinogen concentrations: early substitution and conventional substitution. In both surgical settings, participants in the conventional substitution arm received a greater quantity of red cells within the first 24 hours following surgery than participants in the early substitution arm (see Table 1), although there was no evidence of a statistically significant difference between the intervention arms in either trial (Haas 2015 (C): MD −8.70 (95% CI −21.71 to 4.31) mL/kg; 30 participants; and Haas 2015(S): MD −3.50, 95% CI −9.99 to 2.99 mL/kg; 19 participants; (Analysis 2.5).

All 30 participants in the craniosynostosis trial (Haas 2015 (C)) received a RBC transfusion within 24 hours of surgery. There was no evidence of a difference in the number of participants requiring a transfusion at 24 hours following surgery in the scoliosis trial (Haas 2015(S)): RR 0.60, 95% CI 0.25 to 1.46; 19 participants).

Fibrinogen concentrate: therapeutic trials with inactive comparator

(Eight trials: Bilecen 2017; Collins 2017; Curry 2018; Jeppsson 2016; Nascimento 2016; Rahe‐Meyer 2013; Rahe‐Meyer 2016; Wikkelso 2015)

Six trials including 568 participants reported data for RBC transfusion requirement (Bilecen 2017; Collins 2017; Curry 2018; Jeppsson 2016; Rahe‐Meyer 2013; Wikkelso 2015). All six trials reported the number of participants who received a RBC transfusion, although differences in the measurement time points and the clinical setting of the trials prevented us from pooling the data (Analysis 3.5).

Bilecen 2017 and Rahe‐Meyer 2013 reported the incidence of receipt of a RBC transfusion within 24 hours of delivery of the study drug in the management of bleeding during cardiac surgery. There was evidence of a reduction in receipt of a RBC transfusion for the participants who received fibrinogen in comparison with the placebo group: RR 0.54, 95% CI 0.39 to 0.75; P < 0.001; 178 participants; I² = 0%; Analysis 3.5). The investigators also reported the median number of red blood cell units transfused (within 24 hours of delivery of the study drug), which was less in the fibrinogen group than in the placebo group (see Table 1). This finding was not repeated at the later time point reported by Jeppsson 2016, who found no evidence of a difference in the number of participants requiring a RBC transfusion during their entire hospital stay: RR 1.00, 95% CI 0.38 to 2.66; 48 participants.

Wikkelso 2015 found no evidence of a difference in the receipt of a RBC transfusion in 244 women with a postpartum haemorrhage who received fibrinogen or placebo (saline) within four hours (RR 0.39, 95% CI 0.13 to 1.22), within 24 hours (RR 0.72, 95% CI 0.38 to 1.38), and within seven days (RR 0.95, 95% CI 0.58 to 1.54) of receipt of the study drug. The investigators also reported the number of women who received a transfusion of four units or more (time point unspecified), with no evidence of a difference between the fibrinogen and the placebo (saline) group: RR 2.74, 95% CI 0.71 to 10.57; 244 participants. Similarly Collins 2017 found no evidence of a difference in the number of participants receiving a red blood cell transfusion between the fibrinogen concentrate and placebo arms before hospital discharge for postpartum haemorrhage: RR 0.96, 95% CI 0.55 to 1.69; 55 participants.

Curry 2018 found no difference in the incidence of RBC transfusion following admission for trauma at three hours (RR 1.05, 95% CI 0.86 to 1.29; 46 participants), at six hours (RR 1.05, 95% CI 0.84 to 1.30; 43 participants), or 24 hours (RR 1.11, 95% CI 0.91 to 1.36; 41 participants).

Only Jeppsson 2016 reported the volume of red blood cells transfused; during the hospital stay following cardiac surgery there was no evidence of a difference between participants in whom bleeding was managed with fibrinogen concentrate and those who received placebo: MD −210.00, 95% CI −608.81 to 188.81 mL; 48 participants (Analysis 3.6).

Median values are reported in Table 1. These data were presented in seven studies (Bilecen 2017; Collins 2017; Curry 2018; Jeppsson 2016; Nascimento 2016; Rahe‐Meyer 2013; Rahe‐Meyer 2016).

Fibrinogen concentrate: therapeutic trials with active comparator

(Four trials: Galas 2014; Innerhofer 2017; Lance 2012; Tanaka 2014)

All four trials (total of 220 participants) reported data for RBC transfusion requirement. Lance 2012 reported the volume of RBC transfusion required following major elective surgery when fibrinogen concentrate was used as a partial replacement for FFP in the management of bleeding, but found no evidence of a difference between the intervention arms (MD −120.00, 95% CI −546.93 to 306.93 mL; 43 participants (Analysis 4.5).

Three trials (total of 177 participants) reported the number of participants receiving a RBC transfusion (Galas 2014; Innerhofer 2017; Tanaka 2014), but due to differences in the haemostatically active control used and the time points reported we did not undertake any meta‐analyses. Following cardiac surgery Tanaka 2014 reported no evidence of a difference in red cell exposure by 24 hours post‐operatively between the study's fibrinogen and platelet arms: RR 1.00, 95% CI 0.75 to 1.34; 20 participants. Innerhofer 2017 reported no evidence of a difference between fibrinogen and FFP for red cell exposure 24 hours after a trauma admission: RR 1.04, 95% CI 0.90 to 1.21; 94 participants, and Galas 2014 found no difference at post‐operative day seven between fibrinogen and cryoprecipitate: RR 0.86, 95% CI 0.72 to 1.02; 63 participants (Analysis 4.6).

Median values are reported in Table 1 for Innerhofer 2017.

Factor XIII (FXIII): prophylactic trials with inactive comparator

(Six trials: Godje 2006; Karkouti 2013; Korte 2009; Levy 2009; Rasche 1982; Shirahata 1990)

Five trials (472 participants) reported data for RBC transfusion requirement (Godje 2006; Karkouti 2013; Korte 2009; Levy 2009; Rasche 1982).

One study (Godje 2006) reported the volume of red blood cells transfused and found no evidence of a difference between the intervention arms: −MD 0.70, 95% CI −1.63 to 0.23; 50 participants; Analysis 5.5). Karkouti 2013 reported data for the number of participants who received a RBC transfusion following elective surgery; there was no evidence of a difference between factor XIII and placebo: RR 0.97, 95% CI 0.70 to 1.35; 266 participants (Analysis 5.6). Levy 2009 stated that this outcome was measured in their trial, but did not report their findings.

Data were reported as medians with IQRs only for Karkouti 2013, Korte 2009 and Rasche 1982. In Korte 2009, the surgical participants in the haemostatically inactive control (placebo) arm received more RBC units than participants in the haemostatically active arm. In Rasche 1982, people with acute leukaemia in the haemostatically active arm required more RBC units than those in the haemostatically inactive control (placebo) arm. Median values are reported in Table 1.

Factor XIII (FXIII):therapeutic trials with inactive comparator

(One trial: Bregenzer 1999)

Bregenzer 1999 did not report on this outcome.

Prothrombin complex concentrate (PCC): prophylactic trials with inactive comparator

(One trial: Turner 1981)

Turner 1981 did not report on this outcome.

Blood loss

See Table 2 for individual study data for this outcome, which was reported in a number of ways (means, median, total) across the included studies.

Fibrinogen concentrate: prophylactic trials with inactive comparator

(Nine trials: Cui 2010; Fenger Erikson 2009; Karlsson 2011; Najafi 2014; Pournajafian 2015; Ranucci 2015; Sabate 2016; Sadeghi 2014; Soleimani 2016)

Post‐operative bleeding was reported by eight trials with 378 participants (Cui 2010; Fenger Erikson 2009; Karlsson 2011; Najafi 2014; Pournajafian 2015; Ranucci 2015, Sadeghi 2014; Soleimani 2016) (Analysis 1.7).

In three trials, there was evidence of a difference in blood loss between the treatment arms with more blood being lost in the inactive comparator group than in the group that received fibrinogen concentrate (Analysis 1.7). We pooled data for the two cardiac surgery trials which reported blood loss at 12 hours post‐operatively (Karlsson 2011; Sadeghi 2014): Ratio of Geometric Means 0.68, 95% CI 0.60 to 0.76; P < 0.001; I² = 0%; 180 participants. Pournajafian 2015 reported evidence of a 21% reduction in bleeding at 24 hours following spinal fusion surgery for prophylactic fibrinogen concentration in comparison to an inactive comparator: Ratio of Geometric Means 0.77, 95% CI 0.66 to 0.89; P < 0.001; 41 participants.

There was no evidence of a difference between the intervention arms:

in intra‐operative (Ratio of Geometric Means 0.90, 95% CI 0.64 to 1.27; 30 participants) or 24‐hour blood loss during hip surgery (Ratio of Geometric Means 0.99, 95% CI 0.80 to 1.24; 30 participants) in Najafi 2014;
in intra‐operative (Ratio of Geometric Means 1.02, 95% CI 0.75 to 1.37; 60 participants); or 48‐hour blood loss following prostatic resection (Ratio of Geometric Means 0.85, 95% CI 0.57 to 1.27; 60 participants) in Soleimani 2016;
in blood loss at one hour following cardiac surgery (Ratio of Geometric Means 0.75, 95% CI 0.52 to 1.09; 31 participants) in Cui 2010;
in blood loss during and up to 48 hours after radical cystectomy (Ratio of Geometric Means 0.94, 95% CI 0.67 to 1.33; 20 participants) in Fenger Erikson 2009.

Ranucci 2015 reported the median quantity of blood lost from chest drain in the first 12 hours following elective cardiac surgery in 116 participants: there was less blood loss in participants receiving fibrinogen compared to the placebo control (saline); see Table 2.

Fibrinogen concentrate: prophylactic trials with active comparator

(Two participant groups reported in one trial; Haas 2015 (C); Haas 2015 (S))

Haas 2015 (C) and Haas 2015 (S) reported the median calculated total blood loss at 24 hours following surgery. This trial compared receipt of fibrinogen concentrate at two predefined intraoperative fibrinogen concentrations: early substitution and conventional substitution. In both surgical settings, participants in the conventional substitution arm had a greater total blood loss 24 hours post‐operatively than those in the early substitution arm. Median values are reported in Table 2.

Fibrinogen concentrate: therapeutic trials with inactive comparator

(Eight trials: Bilecen 2017; Collins 2017; Curry 2018; Jeppsson 2016; Nascimento 2016; Rahe‐Meyer 2013; Rahe‐Meyer 2016; Wikkelso 2015)

Five trials with 614 participants reported data for blood loss post‐intervention.

The only trial that reported mean blood loss following use of fibrinogen concentrate as a treatment for bleeding (Jeppsson 2016) found no evidence of a difference in bleeding volume during surgery (Ratio of Geometric Mean 0.87, 95% CI 0.68 to 1.11; 48 participants) or at 12 hours post‐operatively (Ratio of Geometric Mean 0.86, 95% CI 0.67 to 1.11; 48 participants), although blood loss was lower in the fibrinogen arm at all time points (Analysis 3.7).

Median blood loss was reported in five trials (Bilecen 2017; Collins 2017; Jeppsson 2016; Rahe‐Meyer 2016; Wikkelso 2015). Bilecen 2017 provided median blood loss data at seven time points following cardiac surgery up to 24 hours post‐operatively. Participants who received fibrinogen demonstrated lower bleeding volumes at all time points. Collins 2017 and Wikkelso 2015 provided median haemorrhage volumes following postpartum haemorrhage. In the fibrinogen intervention arm, Collins 2017 found there was less blood loss at 24 hours following study intervention, while in Wikkelso 2015 there was no difference at up to six weeks postpartum. Jeppsson 2016 and Rahe‐Meyer 2016 reported median blood loss after cardiac surgery. Jeppsson 2016 found comparable bleeding volumes during surgery with lower volumes at 12 hours post‐operatively in the fibrinogen arm. Rahe‐Meyer 2016 demonstrated lower blood loss at 6, 12 and 24 hours post‐operatively in the active intervention arm.

Median values are reported in Table 2.

Fibrinogen concentrate: therapeutic trials with active comparator

(Four trials: Galas 2014; Innerhofer 2017; Lance 2012; Tanaka 2014)

Two trials with 80 participants reported median data for blood loss post‐intervention (Galas 2014; Tanaka 2014). Tanaka 2014 reported median blood loss as chest drainage to 12 hours post‐cardiac surgery, and Galas 2014 reported median blood loss at 48 hours following cardiac surgery. In both trials those participants who received the fibrinogen concentrate lost less blood than participants who received the respective haemostatically active control: cryoprecipitate in Galas 2014, and platelets in Tanaka 2014. Median values are reported in Table 2.

Factor XIII (FXIII): prophylactic trials with inactive comparator

(Six trials: Godje 2006; Karkouti 2013; Korte 2009; Levy 2009; Rasche 1982; Shirahata 1990)

Only two trials (total of 66 participants) reported mean blood loss following prophylactic administration of factor XIII, both in cardiac surgery (Godje 2006; Levy 2009). Blood loss was reported at a range of time points; at all time points blood loss was lower in the haemostatically active arm. We pooled data for two common time points, eight and 24 hours post‐operatively. There was evidence of reduced blood loss in participants who received factor XIII when compared to an inactive comparator at eight hours (Ratio of Geometric Mean 0.79, 95% CI 0.65 to 0.97; P = 0.02; I² = 0%; 66 participants) but this did not persist at 24 hours post‐operatively (Ratio of Geometric Mean 0.85, 95% CI 0.72 to 1.00; 66 participants) (Analysis 5.7).

Godje 2006 reported data at three further time points, with evidence of a reduction in blood loss for participants who had received factor XIII, in comparison to placebo at 12 hours (Ratio of Geometric Mean 0.78, 95% CI 0.63 to 0.97; P = 0.02; 50 participants), at 36 hours (Ratio of Geometric Mean 0.82, 95% CI 0.68 to 0.98; P = 0.03; 50 participants), and at 48 hours (Ratio of Geometric Mean 0.73, 95% CI 0.62 to 0.86; P < 0.001; 50 participants) post‐operatively. Levy 2009 reported blood loss data at the time of drain removal and observed that there was no evidence of a difference in blood loss between the intervention arms at those time points (Ratio of Geometric Mean 0.68, 95% CI 0.32 to 1.47; 16 participants).

Korte 2009 reported median blood loss at the completion of surgery for gastrointestinal cancer and noted greater blood loss in the Factor XIII arm when compared to placebo, see Table 2 for median values.

Factor XIII (FXIII): therapeutic trials with inactive comparator

(One trial: Bregenzer 1999)

Bregenzer 1999 did not report on blood loss post‐intervention.

Prothrombin complex concentrate (PCC): prophylactic trials with inactive comparator

(One trial: Turner 1981)

Turner 1981 did not report on blood loss post‐intervention.

Secondary outcome: allergic adverse reactions

Fibrinogen concentrate: prophylactic trials with inactive comparator

(Nine trials: Cui 2010; Fenger Erikson 2009; Karlsson 2011; Najafi 2014; Pournajafian 2015; Ranucci 2015; Sabate 2016; Sadeghi 2014; Soleimani 2016)

One cardiac surgery trial (Ranucci 2015), and two non‐cardiac surgery trials (Fenger Erikson 2009; Soleimani 2016), with 196 participants in total, provided data on the incidence of allergic or hypersensitivity events following fibrinogen concentrate administration. No adverse events were reported at any time point.

Fibrinogen concentrate: prophylactic trials with active comparator

(Two participant groups reported in one trial: Haas 2015 (C); Haas 2015 (S))

No allergic adverse events to fibrinogen concentrate were reported in either arm of Haas 2015 (C) or Haas 2015 (S) in a total of 49 participants up to 24 hours post‐operatively.

Fibrinogen concentrate: therapeutic trials with inactive comparator

(Eight trials; Bilecen 2017; Collins 2017; Curry 2018; Jeppsson 2016; Nascimento 2016; Rahe‐Meyer 2013; Rahe‐Meyer 2016; Wikkelso 2015)

Four trials reported data on hypersensitivity reactions to fibrinogen concentrate or placebo in 470 participants. A single allergic event was recorded within 24 hours of placebo administration in the setting of postpartum haemorrhage (Wikkelso 2015: RR 0.33 (95% CI 0.01 to 8.06); 244 participants). No allergic reactions were reported at any time point following cardiac surgery (Bilecen 2017; Rahe‐Meyer 2013), or trauma (Nascimento 2016).

Fibrinogen concentrate: therapeutic trials with active comparator

(Four trials: Galas 2014; Innerhofer 2017; Lance 2012; Tanaka 2014)

Two trials with 106 participants reported data on allergic adverse reactions and found no hypersensitivity reactions to fibrinogen concentrate or the active comparators cryoprecipitate (Galas 2014), and FFP (Lance 2012).

Factor XIII (FXIII): prophylactic trials with inactive comparator

(Six trials: Godje 2006; Karkouti 2013; Korte 2009; Levy 2009; Rasche 1982; Shirahata 1990)

Five trials (Godje 2006; Korte 2009; Levy 2009; Rasche 1982; Shirahata 1990) reporting on 206 participants reported no incidents of allergy to factor XIII, although the time points were unclear in all of these studies.

Factor XIII (FXIII): therapeutic trials with inactive comparator

(One trial: Bregenzer 1999)

Bregenzer 1999 found no evidence of hypersensitivity to factor XIII in 28 participants.

Prothrombin complex concentrate (PCC): prophylactic trials with inactive comparator

(One trial: Turner 1981)

Turner 1981 reported no allergic reactions during administration of PCC prophylaxis in 78 participants.

Change to laboratory measures of coagulation, both standard laboratory tests and global measurements of coagulation

Due to the volume and heterogeneity of time points, definitions, reporting formats and tests performed, we did not undertake statistical analyses of these data. Instead we report the number of studies which reported on this outcome and the measures that they used to assess the outcome.

Fibrinogen concentrate: prophylactic trials with inactive comparator

(Nine trials: Cui 2010; Fenger Erikson 2009; Karlsson 2011; Najafi 2014; Pournajafian 2015; Ranucci 2015; Sabate 2016; Sadeghi 2014; Soleimani 2016)

Eight trials (410 participants) reported measures of coagulation (Cui 2010; Fenger Erikson 2009; Karlsson 2011; Najafi 2014; Pournajafian 2015; Ranucci 2015; Sabate 2016; Sadeghi 2014). All studies reported standard laboratory measures (most commonly Clauss fibrinogen levels) except Cui 2010, where only TEG results were reported. Three studies reported viscoelastic tests and standard laboratory measures (Fenger Erikson 2009; Ranucci 2015; Sabate 2016), and only one study reported additional global testing (thrombin generation: Fenger Erikson 2009). All studies took samples at baseline and five reported a coagulation outcome at 24 hours (Karlsson 2011; Najafi 2014; Pournajafian 2015; Sabate 2016; Sadeghi 2014). There was marked heterogeneity across the studies in timing and frequency of testing.

Fibrinogen concentrate: prophylactic trials with active comparator

(Two participant groups reported in one trial: Haas 2015 (C); Haas 2015 (S))

These two trials analysed standard coagulation tests (activated partial thromboplastin time, international normalised ratio (INR), Clauss fibrinogen and factor XIII) and ROTEM at baseline alone (Haas 2015 (C); Haas 2015 (S)).

Fibrinogen concentrate: therapeutic trials with inactive comparator

(Eight trials: Bilecen 2017; Collins 2017; Curry 2018; Jeppsson 2016; Nascimento 2016; Rahe‐Meyer 2013; Rahe‐Meyer 2016; Wikkelso 2015)

All eight trials (with 759 participants) reported standard clotting tests, including Clauss fibrinogen, with all studies reporting results from baseline. Three studies (Collins 2017; Rahe‐Meyer 2013; Rahe‐Meyer 2016) reported additional ROTEM measures, with all three using a FIBTEM test (a ready‐to‐use ROTEMsystem reagent for use with citrated whole blood, and assesses the firmness of the fibrin clot). The reporting of post‐intervention coagulation tests varied, particularly with timing of tests. Between two and 10 tests were performed between two hours and day 11 post‐intervention. The commonest reported time was 24 hours, which was conducted in six studies (Bilecen 2017; Collins 2017; Jeppsson 2016; Nascimento 2016; Rahe‐Meyer 2016; Wikkelso 2015).

Fibrinogen concentrate: therapeutic trials with active comparator

(Four trials: Galas 2014; Innerhofer 2017; Lance 2012; Tanaka 2014)

All four trials (with 220 participants) reported standard clotting assays and viscoelastic measures (all used ROTEM). Three studies reported coagulation factor levels, which varied between trials: factors II, VII, IX, X and XIII (Galas 2014); factors II, VIII, IX, X, anti‐thrombin (Lance 2012); and factor XIII and anti‐thrombin (Innerhofer 2017). All studies reported baseline results and three of the four reported 24‐hour results (Galas 2014; Innerhofer 2017; Tanaka 2014).

Factor XIII (FXIII): prophylactic trials with inactive comparator

(Six trials: Godje 2006; Karkouti 2013; Korte 2009; Levy 2009; Rasche 1982; Shirahata 1990)

Five trials reported FXIII levels at baseline and at one or more subsequent time points (Godje 2006; Karkouti 2013; Korte 2009; Levy 2009; Rasche 1982), which varied widely between 30 minutes post‐intervention (Karkouti 2013; Levy 2009) to once a week testing (Rasche 1982). One trial, with 58 neonates, reported only one set of coagulation tests (factor XIII, Clauss fibrinogen, plasminogen, anti‐thrombin, proteins C and S, alpha‐2 anti‐plasmin, plasmin‐anti‐plasmin and fibrinogen degradation products), taken within the first week of birth (Shirahata 1990). Standard clotting tests (activated partial thromboplastin time, prothrombin time and Clauss fibrinogen) were reported in two studies (Godje 2006; Rasche 1982), and two studies reported Clauss fibrinogen only (Korte 2009; Shirahata 1990). Only one study reported ROTEM measures (Korte 2009).

Factor XIII (FXIII): therapeutic trials with inactive comparator

(One trial: Bregenzer 1999)

Bregenzer 1999 reported FXIII measures at nine time points, including baseline and up to day 42 post‐therapy in 20 participants. No other coagulation parameter was reported.

Prothrombin complex concentrate (PCC): prophylactic trials with inactive comparator

(One trial: Turner 1981)

Standard clotting assays (activated partial thromboplastin time, prothrombin time, Clauss fibrinogen) were reported at two time points, baseline and post‐intervention, in the 78 neonates in this trial (Turner 1981).

Discussion

Summary of main results

We identified 31 randomised controlled trials which involved a total of 2392 participants, and compared the use of intravenous pro‐coagulant haemostatic factors with either placebo or another haemostatically active agent.

Unlike inherited bleeding conditions, where a person has a deficiency in one coagulation factor alone, acquired coagulopathy is more complex, often involving multiple deficiencies of coagulant factors, and which also evolves with time. For example, in trauma haemorrhage, when trauma coagulopathy is present it leads initially to a reduction in pro‐coagulant factors (i.e. including factors V, XIII and fibrinogen); an increase in fibrinolytic factors (i.e. tissue plasminogen activator (t‐PA) with low plasminogen activator inhibitor type 1 (PAI‐1) and alpha 2‐antiplasmin (A2‐AP)) as well as an alteration of anticoagulant factors (i.e. high activated protein C levels). These initial changes can be associated with a bleeding phenotype, although at later stages can shift towards a pro‐thrombotic state. This has implications for the management of acquired bleeding, as it suggests that single pro‐haemostatic therapies may have less effect. This review therefore was aimed at understanding the trial data to inform efficacy and safety of multiple pro‐coagulants, or in combination.

Our review identified data on three different pro‐haemostatic agents as interventions: fibrinogen (23 trials), Factor XIII (7 trials) and PCC (1 trial). While the earlier studies focused on factor XIII, there has been an increase in the investigation of fibrinogen concentrate since 2010, as demonstrated by fibrinogen concentrate being the focus of most of the trials published in the last five years. Eighteen trials explored the use of these pro‐haemostatic agents as prophylaxis for bleeding and 13 trials explored their therapeutic use. Fibrinogen was the focus of 12 of the 13 therapy‐orientated trials whilst six of the factor XIII trials tested used it as a prophylactic agent.

Summary of primary outcomes

The primary outcomes were all‐cause mortality (reported by 21 trials), arterial thromboembolic events (reported by 22 trials) and venous thromboembolic events (reported by 21 trials). The results of meta‐analyses did not demonstrate an impact of pro‐haemostatic agents on mortality or the incidence of arterial or venous thromboembolic events. The overall quality of this evidence ranged from very low to high (summary of findings Table for the main comparison; summary of findings Table 2; summary of findings Table 3; summary of findings Table 4; summary of findings Table 5).

Summary of secondary outcomes

Data were available for all of our secondary outcomes: mortality due to bleeding, red blood cell transfusion requirement, blood loss and allergic reaction. We also reported on changes to laboratory measures of coagulation, both standard laboratory tests and global measurements of coagulation, but meta‐analysis of these data were limited by heterogeneity in the assays and reported results.

Overall there was no evidence of a decrease in mortality due to bleeding in any of the 10 trials that contributed data for this outcome. Fibrinogen concentrate was associated with a reduction in the individual patient risk of being exposed to a red blood cell transfusion when compared to inactive comparators. This finding was evident in both cardiac and non‐cardiac surgery settings and in the prophylactic and therapeutic settings (Analysis 1.6, Analysis 3.5). Additionally, prophylactic fibrinogen concentrate was found to reduce post‐operative bleeding in two studies of cardiac surgery and one trial in orthopaedic surgery, when compared to a haemostatically inactive comparator (Analysis 1.7). Overall, the prophylactic use of fibrinogen concentrate almost halves the risk of a person requiring a blood transfusion following cardiac surgery and reduces it by 75% in non‐cardiac surgery compared to placebo (Analysis 1.6). Prophylactic factor XIII demonstrated efficacy at reducing the volume of post‐operative bleeding in cardiac surgery compared to placebo (Analysis 5.7).

All studies that reported blood loss following surgery found greater blood loss in the haemostatically inactive arm (no use of pro‐haemostatic agents), irrespective of differences in the clinical setting, the intervention or how and when blood loss was measured. Across all trials only a single intervention‐related allergic adverse event was recorded in an inactive‐control study arm participant.

Overall the data evaluating coagulation test results were highly variable, with data presenting a wide range of standard as well as global tests (viscoelastic and thrombin generation). There was no consensus on the timing or frequency of results which were used to evaluate the 'laboratory efficacy' of the intervention in each clinical trial. Of more importance is the overwhelming lack of any attempt by trialists to combine laboratory data with clinically meaningful endpoints; for example, if a pro‐haemostatic drug is being given with the aim of preventing or treating bleeding, laboratory data would be best presented in combination with clinical endpoints such as 'cessation of bleeding' or 'time to haemostasis'.

Overall completeness and applicability of evidence

Completeness

The interpretation of the findings of our review are limited by the individual trial data that contributed to meta‐analyses. Specifically, meta‐analysis of the data across trials was not possible for many outcomes, due to a lack of common time points or inadequate or inconsistent definitions of time points. We were only able to report data on our primary outcomes of mortality and thromboembolic events in 21 out of 31 trials.

The identified studies recruited participants across a broad range of clinical settings in which pro‐haemostatic agents are likely to be used. These include trauma, postpartum haemorrhage and cardiac surgery, where clinical outcomes might be modified significantly by optimal use of pro‐haemostatic agents. However, other clinical areas of bleeding risk, such as gastro‐intestinal bleeding, were poorly represented in the included trials.

Applicability

Applicability of the review findings is limited. None of the included studies was powered to detect harm or adverse events, and there is little documentation of adverse events to the 90‐day time point that is recommended for hospital‐associated thromboembolic complications. The poor quality of the evidence means that the results for outcomes assessed in this review cannot be applied in other clinical settings (See 'Summary of findings' tables).

Quality of the evidence

We used GRADE to summarise the quality of evidence for the two primary outcomes: mortality and thromboembolic events in the 'Summary of findings' tables. We focused on mortality for the studies assessing the use of pro‐coagulant haemostatic agents as a treatment for bleeding and thromboembolic events to assess the prophylactic use of pro‐coagulant haemostatic agents. The overall quality of the evidence ranged from very low to high, with most trial outcomes being rated as low quality. We considered no trial to be at low risk of bias in all domains, but we downgraded half the outcomes by one level for risk of bias. Domains with high risk of bias included allocation concealment, blinding of study personnel and outcome assessors, incomplete outcome data and selective reporting. The small cohorts and rare mortality and thrombotic events introduced risks of imprecision. Lastly, the trials in this review represented most of the clinical areas in which bleeding is observed, but not all clinical areas were represented in each of the intervention comparisons. Moreover the trials did not set out to explore the outcomes of interest to this review, and this introduced inconsistency. Lastly, we did not downgrade any outcome due to inconsistency, perhaps as a consequence of GRADE assessments being based on outcome data for one or two studies due to diversity in the clinical setting and the timing of outcome measurements, which limited our ability to pool data across studies (as has been highlighted above).

Potential biases in the review process

The searches performed for this review were extensive and we have tried to identify all trials eligible for inclusion, regardless of their publication status or language of publication. We have also attempted to obtain missing data from all the included studies to provide a complete picture of the primary and secondary review outcomes, and were able to acquire this from eight authors. The review authors performed screening and data extraction in duplicate to minimise bias and were blinded to each other's results.

Included study bias

Determining the risks of bias of these studies highlighted a range of methodological issues. A significant number of studies were unclear in their randomisation methodology (15 studies) and allocation concealment (14 at unclear risk and five at high risk of bias). Any impairment of the randomisation process is known to generate an overestimation of the clinical efficacy of any intervention (Schulz 1995). Similarly the blinding status of outcome assessors was often not addressed (13 at unclear risk and five at high risk of bias), although most outcomes in these trials were objective and therefore not prone to observer bias. More important, however, is the frequency with which study personnel were either unblinded or where insufficient information was available to assess their level of blinding (five at unclear risk and seven at high risk of bias). This is concerning for outcomes such as transfusion, especially where no algorithm or threshold for transfusion was applied, where decisions are likely to be prone to bias.

Limitations of this review

Most of the studies we identified that contributed to this review were small‐scale, with a mean of 78 participants (range 20 to 479). This review is clearly limited to the data published and provided by the trial authors. A surprising number of studies did not provide mortality data as an outcome (10 studies). Most outcomes of interest in this review are objective and have well‐defined criteria, although the definitions of thromboembolic events, allergy and blood loss provided were less robust.

Sample sizes in future randomized trials powered on clinical outcomes of mortality (or even preferably mortality due to bleeding) would need to be greatly increased. Analogous examples for power calculations would be CRASH2 (CRASH‐2 2010), WOMAN (WOMAN 2017) and HALT IT (Roberts 2014, Brenner 2018), which recruited between 15,000 and 20,000 participants each.

We have grouped the wide range of study clinical settings into the broad primary categories of therapy or prophylaxis. However, we recognise heterogeneity within these settings and at further levels. For example, trials within the cardiac group may include more routine coronary artery bypass graft surgery (Karlsson 2011; Sadeghi 2014), or complex aortic surgery (Rahe‐Meyer 2016; Ranucci 2015), presumably with different levels of bleeding risk.

The study interventions, doses and administration regimens varied. However, several of these agents act through common pathways of coagulation (factor XIII and fibrinogen, for example, principally act to increase clot strength and reduce the susceptibility of clots to fibrinolysis), and in some recent trials these interventions were used in combination (Innerhofer 2017). Fibrinogen concentrate doses ranged from one gram (Pournajafian 2015; Sadeghi 2014) to six grams (Curry 2018; Nascimento 2016), and factor XIII therapy ranged from a single dose (Godje 2006; Karkouti 2013) to nine daily doses (Bregenzer 1999). In this review we have not differentiated between the doses given, as at present there are no robust data to support any one dose or dose calculation as providing clinical efficacy in a given setting. However, the more recent fibrinogen concentrate papers have implemented fibrinogen‐dosing protocols based on fibrinogen levels or viscoelastic markers of clot strength (Bilecen 2017; Collins 2017; Cui 2010; Haas 2015; Rahe‐Meyer 2013; Ranucci 2015; Sabate 2016). While this approach might eventually permit individualised approaches to product dosing, it increases the difficulty of comparing these studies. Additionally it is likely that some studies reported here administered haemostatic agent doses below that required for clinical efficacy (Pournajafian 2015; Sadeghi 2014), with the potential of false negative findings.

The published data on markers and indices of coagulation status were very heterogeneous and it is difficult to offer a useful interpretation of these data.

Agreements and disagreements with other studies or reviews

There are no other published reviews of this scope on factor XIII for comparison. The paucity of data for the use of pro‐thrombin complex concentrates may have precluded systematic reviews of randomised trials, although there are reviews of observational studies (Zeng 2017).

Clinical findings and conclusions

The findings of this review are consistent with other systematic reviews of the use of fibrinogen concentrate. An analysis of one prospective and seven retrospective studies in trauma found no impact on overall mortality and recorded a broad range of recorded mortality time points and heterogeneity in the reported outcomes (Mengoli 2017). Similarly, analysis of blood loss and transfusion was impaired by variation in the use of continuous and discrete data presentation. Another systematic review that included four studies (two RCTs and two non‐randomised controlled studies) concluded that fibrinogen concentrate in the perioperative and major haemorrhage settings was safe and reduced bleeding and the transfusion of blood components (Franchini 2012). A Cochrane Review of fibrinogen concentrate in bleeding patients was published in 2013 and contains six RCTs that compared fibrinogen to placebo or usual care in adults with bleeding (Wikkelso 2013). We have included these RCTs in this review (Cui 2010; Fenger Erikson 2009; Galas 2014; Karlsson 2011; Lance 2012; Rahe‐Meyer 2013). In keeping with our conclusions, the authors found that fibrinogen concentrate reduced transfusion requirement, but there were insufficient data to draw conclusions about the impact on mortality or other clinical efficacy.

Risk of bias and quality of included studies

The three reviews highlighted above reached comparable conclusions to this review about the poor quality of evidence and frequently indeterminate or high risks of bias in the included studies.

Figure 1

Trial flow diagram.

Ir a la figura de la revisiónAbrir en una pestaña nueva

Figure 2

Risk of bias summary: review authors' judgements about each risk of bias item for each included trial. Thirty‐one trials are included in this review (Haas 2015 reports two trials).

Ir a la figura de la revisiónAbrir en una pestaña nueva

Analysis 1.1

Comparison 1 Fibrinogen: prophylactic trials: intervention vs inactive control, Outcome 1 All‐cause mortality.

Ir a la figura de la revisiónAbrir en una pestaña nueva

Analysis 1.2

Comparison 1 Fibrinogen: prophylactic trials: intervention vs inactive control, Outcome 2 Arterial thromboembolic events.

Ir a la figura de la revisiónAbrir en una pestaña nueva

Analysis 1.3

Comparison 1 Fibrinogen: prophylactic trials: intervention vs inactive control, Outcome 3 Venous thromboembolic events.

Ir a la figura de la revisiónAbrir en una pestaña nueva

Analysis 1.4

Comparison 1 Fibrinogen: prophylactic trials: intervention vs inactive control, Outcome 4 Mortality due to bleeding.

Ir a la figura de la revisiónAbrir en una pestaña nueva

Analysis 1.5

Comparison 1 Fibrinogen: prophylactic trials: intervention vs inactive control, Outcome 5 RBC transfusion requirement.

Ir a la figura de la revisiónAbrir en una pestaña nueva

Analysis 1.6

Comparison 1 Fibrinogen: prophylactic trials: intervention vs inactive control, Outcome 6 Receipt of a RBC transfusion.

Ir a la figura de la revisiónAbrir en una pestaña nueva

Analysis 1.7

Comparison 1 Fibrinogen: prophylactic trials: intervention vs inactive control, Outcome 7 Blood loss.

Ir a la figura de la revisiónAbrir en una pestaña nueva

Analysis 1.8

Comparison 1 Fibrinogen: prophylactic trials: intervention vs inactive control, Outcome 8 Allergic Adverse Events.

Ir a la figura de la revisiónAbrir en una pestaña nueva

Analysis 2.1

Comparison 2 Fibrinogen: prophylactic trials: intervention vs haemostatically active control, Outcome 1 All‐cause mortality.

Ir a la figura de la revisiónAbrir en una pestaña nueva

Analysis 2.2

Comparison 2 Fibrinogen: prophylactic trials: intervention vs haemostatically active control, Outcome 2 Arterial thromboembolic events.

Ir a la figura de la revisiónAbrir en una pestaña nueva

Analysis 2.3

Comparison 2 Fibrinogen: prophylactic trials: intervention vs haemostatically active control, Outcome 3 Venous thromboembolic events.

Ir a la figura de la revisiónAbrir en una pestaña nueva

Analysis 2.4

Comparison 2 Fibrinogen: prophylactic trials: intervention vs haemostatically active control, Outcome 4 Mortality due to bleeding.

Ir a la figura de la revisiónAbrir en una pestaña nueva

Analysis 2.5

Comparison 2 Fibrinogen: prophylactic trials: intervention vs haemostatically active control, Outcome 5 RBC transfusion requirement.

Ir a la figura de la revisiónAbrir en una pestaña nueva

Analysis 2.6

Comparison 2 Fibrinogen: prophylactic trials: intervention vs haemostatically active control, Outcome 6 Receipt of a RBC transfusion.

Ir a la figura de la revisiónAbrir en una pestaña nueva

Analysis 2.7

Comparison 2 Fibrinogen: prophylactic trials: intervention vs haemostatically active control, Outcome 7 Allergic Adverse Events.

Ir a la figura de la revisiónAbrir en una pestaña nueva

Analysis 3.1

Comparison 3 Fibrinogen: therapeutic trials: intervention vs inactive control, Outcome 1 All‐cause mortality.

Ir a la figura de la revisiónAbrir en una pestaña nueva

Analysis 3.2

Comparison 3 Fibrinogen: therapeutic trials: intervention vs inactive control, Outcome 2 Arterial thromboembolic events.

Ir a la figura de la revisiónAbrir en una pestaña nueva

Analysis 3.3

Comparison 3 Fibrinogen: therapeutic trials: intervention vs inactive control, Outcome 3 Venous thromboembolic events.

Ir a la figura de la revisiónAbrir en una pestaña nueva

Analysis 3.4

Comparison 3 Fibrinogen: therapeutic trials: intervention vs inactive control, Outcome 4 Mortality due to bleeding.

Ir a la figura de la revisiónAbrir en una pestaña nueva

Analysis 3.5

Comparison 3 Fibrinogen: therapeutic trials: intervention vs inactive control, Outcome 5 Receipt of a RBC transfusion.

Ir a la figura de la revisiónAbrir en una pestaña nueva

Analysis 3.6

Comparison 3 Fibrinogen: therapeutic trials: intervention vs inactive control, Outcome 6 RBC transfusion requirement.

Ir a la figura de la revisiónAbrir en una pestaña nueva

Analysis 3.7

Comparison 3 Fibrinogen: therapeutic trials: intervention vs inactive control, Outcome 7 Blood loss.

Ir a la figura de la revisiónAbrir en una pestaña nueva

Analysis 3.8

Comparison 3 Fibrinogen: therapeutic trials: intervention vs inactive control, Outcome 8 Allergic Adverse Events.

Ir a la figura de la revisiónAbrir en una pestaña nueva

Analysis 4.1

Comparison 4 Fibrinogen: therapeutic trials: intervention vs haemostatically active control, Outcome 1 All‐cause mortality.

Ir a la figura de la revisiónAbrir en una pestaña nueva

Analysis 4.2

Comparison 4 Fibrinogen: therapeutic trials: intervention vs haemostatically active control, Outcome 2 Arterial thromboembolic events.

Ir a la figura de la revisiónAbrir en una pestaña nueva

Analysis 4.3

Comparison 4 Fibrinogen: therapeutic trials: intervention vs haemostatically active control, Outcome 3 Venous thromboembolic events.

Ir a la figura de la revisiónAbrir en una pestaña nueva

Analysis 4.4

Comparison 4 Fibrinogen: therapeutic trials: intervention vs haemostatically active control, Outcome 4 Mortality due to bleeding.

Ir a la figura de la revisiónAbrir en una pestaña nueva

Analysis 4.5

Comparison 4 Fibrinogen: therapeutic trials: intervention vs haemostatically active control, Outcome 5 RBC transfusion requirement.

Ir a la figura de la revisiónAbrir en una pestaña nueva

Analysis 4.6

Comparison 4 Fibrinogen: therapeutic trials: intervention vs haemostatically active control, Outcome 6 Receipt of a RBC transfusion.

Ir a la figura de la revisiónAbrir en una pestaña nueva

Analysis 4.7

Comparison 4 Fibrinogen: therapeutic trials: intervention vs haemostatically active control, Outcome 7 Allergic Adverse Events.

Ir a la figura de la revisiónAbrir en una pestaña nueva

Analysis 5.1

Comparison 5 FXIII: prophylactic trials: intervention vs inactive control, Outcome 1 All‐cause mortality.

Ir a la figura de la revisiónAbrir en una pestaña nueva

Analysis 5.2

Comparison 5 FXIII: prophylactic trials: intervention vs inactive control, Outcome 2 Arterial thromboembolic events.

Ir a la figura de la revisiónAbrir en una pestaña nueva

Analysis 5.3

Comparison 5 FXIII: prophylactic trials: intervention vs inactive control, Outcome 3 Venous thromboembolic events.

Ir a la figura de la revisiónAbrir en una pestaña nueva

Analysis 5.4

Comparison 5 FXIII: prophylactic trials: intervention vs inactive control, Outcome 4 Mortality due to bleeding.

Ir a la figura de la revisiónAbrir en una pestaña nueva

Analysis 5.5

Comparison 5 FXIII: prophylactic trials: intervention vs inactive control, Outcome 5 RBC transfusion requirement.

Ir a la figura de la revisiónAbrir en una pestaña nueva

Analysis 5.6

Comparison 5 FXIII: prophylactic trials: intervention vs inactive control, Outcome 6 Receipt of a RBC transfusion.

Ir a la figura de la revisiónAbrir en una pestaña nueva

Analysis 5.7

Comparison 5 FXIII: prophylactic trials: intervention vs inactive control, Outcome 7 Blood loss.

Ir a la figura de la revisiónAbrir en una pestaña nueva

Analysis 5.8

Comparison 5 FXIII: prophylactic trials: intervention vs inactive control, Outcome 8 Allergic Adverse Events.

Ir a la figura de la revisiónAbrir en una pestaña nueva

Analysis 6.1

Comparison 6 FXIII: therapeutic trials: intervention vs inactive control, Outcome 1 Allergic Adverse Events.

Ir a la figura de la revisiónAbrir en una pestaña nueva

Analysis 7.1

Comparison 7 PCC: prophylactic trials: intervention vs inactive control, Outcome 1 All‐cause mortality.

Ir a la figura de la revisiónAbrir en una pestaña nueva

Analysis 7.2

Comparison 7 PCC: prophylactic trials: intervention vs inactive control, Outcome 2 Allergic Adverse Events.

Ir a la figura de la revisiónAbrir en una pestaña nueva

Summary of findings for the main comparison. FIBRINOGEN: Therapeutic trials: intervention compared to inactive Control for the prevention and treatment of bleeding in participants without haemophilia

FIBRINOGEN: Therapeutic trials: intervention compared to inactive control for the prevention and treatment of bleeding in participants without haemophilia
Patient or population: the prevention and treatment of bleeding in patients without haemophilia Setting: hospital Intervention: FIBRINOGEN in therapeutic trials Comparison: Inactive control
Outcomes	*Anticipated absolute effects^ (95% CI)**		Relative effect (95% CI)	№ of participants (studies)	Certainty of the evidence (GRADE)	Comments
	Risk with inactive control	Risk with fibrinogen in therapeutic trials
All‐cause mortality Up to 28 days following admission	Study population		RR 1.46 (0.71 to 2.99)	97 (2 RCTs)	⊕⊕⊕⊝ MODERATE^a	‐
	184 per 1000	268 per 1000 (130 to 549)
All‐cause mortality In‐hospital mortality up to 30 days post‐operatively	Study population		RR 5.00 (0.25 to 102.00)	120 (1 RCT)	⊕⊕⊝⊝ LOW^a,b	‐
	0 per 1000	0 per 1000 (0 to 0)
All‐cause mortality Up to 6 weeks post‐natally	Study population		‐	294 (2 RCTs)	⊕⊕⊝⊝ LOW^a,b	No events
	see comment	see comment
All‐cause mortality Up to 46 days post‐operatively	Study population		RR 0.23 (0.05 to 1.01)	213 (2 RCTs)	⊕⊕⊕⊕ HIGH	‐
	85 per 1000	20 per 1000 (4 to 86)
*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). CI: Confidence interval; RR: Risk ratio
GRADE Working Group grades of evidence High certainty: We are very confident that the true effect lies close to that of the estimate of the effect Moderate certainty: We are moderately confident in the effect estimate: The true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different Low certainty: Our confidence in the effect estimate is limited: The true effect may be substantially different from the estimate of the effect Very low certainty: We have very little confidence in the effect estimate: The true effect is likely to be substantially different from the estimate of effect
^aWe downgraded by one level as neither study was designed to measure mortality. ^bWe downgraded by one level, due to low event rate.

Summary of findings for the main comparison. FIBRINOGEN: Therapeutic trials: intervention compared to inactive Control for the prevention and treatment of bleeding in participants without haemophilia

Ir a la tabla de la revisión

Summary of findings 2. FIBRINOGEN: Therapeutic trials: intervention compared to haemostatically active control for the prevention and treatment of bleeding in participants without haemophilia

FIBRINOGEN: Therapeutic trials: intervention compared to haemostatically active control for the prevention and treatment of bleeding in participants without haemophilia
Patient or population: the prevention and treatment of bleeding in patients without haemophilia Setting: hospital Intervention: FIBRINOGEN in therapeutic trials Comparison: Haemostatically active control
Outcomes	*Anticipated absolute effects^ (95% CI)**		Relative effect (95% CI)	№ of participants (studies)	Certainty of the evidence (GRADE)	Comments
	Risk with haemostatically active control	Risk with FIBRINOGEN in therapeutic trials
All‐cause mortality Up to 7 days post‐operatively or hospital discharge	Study population		not estimable	63 (1 RCT)	⊕⊕⊕⊝ MODERATE^a	No events
	0 per 1000	0 per 1000 (0 to 0)
All‐cause mortality Up to 28 days post‐operatively	Study population		not estimable	20 (1 RCT)	⊕⊕⊝⊝ LOW^a,b	No events
	0 per 1000	0 per 1000 (0 to 0)
All‐cause mortality Up to 30 days post‐operatively	Study population		RR 0.95 (0.06 to 14.30)	43 (1 RCT)	⊕⊕⊝⊝ LOW^a,c	‐
	48 per 1000	45 per 1000 (3 to 681)
All‐cause mortality Up to 30 days post‐operatively	Study population		RR 2.20 (0.45 to 10.78)	94 (1 RCT)	⊕⊝⊝⊝ VERY LOW^a,c,d	‐
	45 per 1000	100 per 1000 (20 to 490)
*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). CI: Confidence interval; RR: Risk ratio
GRADE Working Group grades of evidence High certainty: We are very confident that the true effect lies close to that of the estimate of the effect Moderate certainty: We are moderately confident in the effect estimate: The true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different Low certainty: Our confidence in the effect estimate is limited: The true effect may be substantially different from the estimate of the effect Very low certainty: We have very little confidence in the effect estimate: The true effect is likely to be substantially different from the estimate of effect
^aWe downgraded by one level due to low event rate. ^bWe downgraded by one level due to concerns about risks of bias both in the way that allocation of treatment was concealed at randomisation and whether blinding to treatment could be maintained for the duration of the trial in this "open‐label trial where adequacy of surgical field haemostasis reported as a subjective outcome". ^cWe downgraded by one level as the study did not measure mortality as an outcome. ^dWe downgraded by one level due to concerns about risks of bias in how blinding to treatment intervention could be maintained for the duration of the trial in this "open‐label trial where adequacy of surgical field haemostasis reported as a subjective outcome", and the non‐reporting of all outcomes that the investigators said they would measure (selective outcome reporting).

Summary of findings 2. FIBRINOGEN: Therapeutic trials: intervention compared to haemostatically active control for the prevention and treatment of bleeding in participants without haemophilia

Ir a la tabla de la revisión

Summary of findings 3. FIBRINOGEN: Prophylactic trials: intervention compared to inactive control for the prevention and treatment of bleeding in participants without haemophilia

FIBRINOGEN: Prophylactic trials: intervention compared to inactive control for the prevention and treatment of bleeding in participants without haemophilia
Patient or population: the prevention and treatment of bleeding in patients without haemophilia Setting: hospital Intervention: FIBRINOGEN: prophylactic trials Comparison: inactive control
Outcomes	*Anticipated absolute effects^ (95% CI)**		Relative effect (95% CI)	№ of participants (studies)	Certainty of the evidence (GRADE)	Comments
	Risk with inactive control	Risk with FIBRINOGEN in prophylactic trials
Venous thromboembolic events Up to 30 days post‐operatively	Study population		‐	122 (2 RCTs)	⊕⊕⊝⊝ LOW^a,b	No events in one trial
	see comment	see comment
Arterial thrombotic events Up to 4 days post‐operatively	Study population		not estimable	80 (2 RCTs)	⊕⊕⊕⊝ MODERATE^a	‐
	25 per 1000	0 per 1000 (0 to 0)
Arterial thrombotic events Up to 30 days post‐operatively	Study population		RR 0.33 (0.01 to 8.02)	116 (1 RCT)	⊕⊕⊝⊝ LOW^a,c	‐
	17 per 1000	6 per 1000 (0 to 138)
Arterial thrombotic events ‐ up to 30 days post‐operatively	Study population		not estimable	122 (2 RCTs)	⊕⊕⊝⊝ LOW^a,b	‐
	51 per 1000	0 per 1000 (0 to 0)
Venous thrombotic events Up to 4 days post‐operatively	Study population		not estimable	80 (2 RCTs)	⊕⊕⊕⊝ MODERATE^a	‐
	0 per 1000	0 per 1000 (0 to 0)
*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). CI: Confidence interval; RR: Risk ratio
GRADE Working Group grades of evidence High certainty: We are very confident that the true effect lies close to that of the estimate of the effect Moderate certainty: We are moderately confident in the effect estimate: The true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different Low certainty: Our confidence in the effect estimate is limited: The true effect may be substantially different from the estimate of the effect Very low certainty: We have very little confidence in the effect estimate: The true effect is likely to be substantially different from the estimate of effect
^aWe downgraded by one level due to low event rate. ^bWe downgraded by one level due to concerns about risks of bias in the non‐reporting of all outcomes that the investigators said they would measure (selective outcome reporting). ^cWe downgraded by one level due to concerns about risks of bias in the way that allocation of the treatment was concealed at the time of randomisation.

Summary of findings 3. FIBRINOGEN: Prophylactic trials: intervention compared to inactive control for the prevention and treatment of bleeding in participants without haemophilia

Ir a la tabla de la revisión

Summary of findings 4. FIBRINOGEN: Prophylactic trials: intervention compared to haemostatically active control for the prevention and treatment of bleeding in participants without haemophilia

FIBRINOGEN: Prophylactic trials: intervention compared to haemostatically active control for the prevention and treatment of bleeding in participants without haemophilia
Patient or population: the prevention and treatment of bleeding in patients without haemophilia Setting: hospital Intervention: FIBRINOGEN: prophylactic trials Comparison: Haemostatically active control
Outcomes	*Anticipated absolute effects^ (95% CI)**		Relative effect (95% CI)	№ of participants (studies)	Certainty of the evidence (GRADE)	Comments
	Risk with haemostatically active control	Risk with FIBRINOGEN in prophylactic trials
Arterial thromboembolic events	Study population		not estimable	49 (1 RCT)	⊕⊕⊝⊝ LOW^a,b	‐
	0 per 1000	0 per 1000 (0 to 0)
Arterial thromboembolic events ‐ Duration of hospital stay (whole population)	Study population		not estimable	49 (1 RCT)	⊕⊕⊝⊝ LOW^a,b	‐
	0 per 1000	0 per 1000 (0 to 0)
Venous thromboembolic events	Study population		not estimable	49 (1 RCT)	⊕⊕⊝⊝ LOW^a,b	‐
	0 per 1000	0 per 1000 (0 to 0)
Venous thromboembolic events Duration of hospital stay (whole population)	Study population		not estimable	49 (1 RCT)	⊕⊕⊝⊝ LOW^a,b	‐
	0 per 1000	0 per 1000 (0 to 0)
*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). CI: Confidence interval; RR: Risk ratio
GRADE Working Group grades of evidence High certainty: We are very confident that the true effect lies close to that of the estimate of the effect Moderate certainty: We are moderately confident in the effect estimate: The true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different Low certainty: Our confidence in the effect estimate is limited: The true effect may be substantially different from the estimate of the effect Very low certainty: We have very little confidence in the effect estimate: The true effect is likely to be substantially different from the estimate of effect
^aWe downgraded by one level as the thromboembolic events (arterial or venous) in these participants given fibrinogen prophylactically were not an outcome of interest. ^bWe downgraded by one level due to the small number of participants available for this outcome.

Summary of findings 4. FIBRINOGEN: Prophylactic trials: intervention compared to haemostatically active control for the prevention and treatment of bleeding in participants without haemophilia

Ir a la tabla de la revisión

Summary of findings 5. FXIII: Prophylactic trials: intervention compared to inactive control for the prevention and treatment of bleeding in participants without haemophilia

FXIII: Prophylactic trials: intervention compared to inactive control for the prevention and treatment of bleeding in participants without haemophilia
Patient or population: the prevention and treatment of bleeding in patients without haemophilia Setting: hospital Intervention: FXIII: prophylactic trials Comparison: inactive control
Outcomes	*Anticipated absolute effects^ (95% CI)**		Relative effect (95% CI)	№ of participants (studies)	Certainty of the evidence (GRADE)	Comments
	Risk with inactive control	Risk with FXIII in prophylactic trials
Arterial thromboembolic events Up to 5 to 7 weeks post‐operatively	Study population		RR 0.67 (0.28 to 1.62)	282 (2 RCTs)	⊕⊕⊝⊝ LOW^a,b	‐
	81 per 1000	54 per 1000 (23 to 131)
Arterial thromboembolic events Up to 30 days post‐operatively	Study population		RR 3.00 (0.14 to 66.53)	22 (1 RCT)	⊕⊕⊝⊝ LOW^b,c	‐
	0 per 1000	0 per 1000 (0 to 0)
Arterial thromboembolic events Post‐operatively	Study population		RR 0.20 (0.01 to 3.97)	50 (1 RCT)	⊕⊝⊝⊝ VERY LOW^b,c,d	‐
	80 per 1000	16 per 1000 (1 to 318)
Venous thromboembolic events ‐ Up to 5 to 7 weeks post‐operatively	Study population		RR 0.93 (0.06 to 14.67)	282 (2 RCTs)	⊕⊕⊝⊝ LOW^a,b	‐
	7 per 1000	7 per 1000 (0 to 108)
Venous thromboembolic events Up to 30 days post‐operatively	Study population		RR 3.00 (0.14 to 66.53)	22 (1 RCT)	⊕⊕⊝⊝ LOW^b,c	‐
	0 per 1000	0 per 1000 (0 to 0)
Venous thromboembolic events Post‐operatively	Study population		not estimable	50 (1 RCT)	⊕⊝⊝⊝ VERY LOW^b,c,d	‐
	0 per 1000	0 per 1000 (0 to 0)
*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). CI: Confidence interval; RR: Risk ratio
GRADE Working Group grades of evidence High certainty: We are very confident that the true effect lies close to that of the estimate of the effect Moderate certainty: We are moderately confident in the effect estimate: The true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different Low certainty: Our confidence in the effect estimate is limited: The true effect may be substantially different from the estimate of the effect Very low certainty: We have very little confidence in the effect estimate: The true effect is likely to be substantially different from the estimate of effect
^aWe downgraded by one level due to concerns about attrition, reporting and selection bias across these trials. ^bWe downgraded by one level as thromboembolic events (arterial or venous) in these participants given factor XIII prophylactically were not an outcome of interest. ^cWe downgraded by one level due to low event rates. ^dWe downgraded by one level due to risks of bias concerns in how concealment of treatment intervention was attempted (unsealed envelopes in the patients notes) and the subsequent impact this had on the level of blinding to treatment intervention that could be maintained for the duration of the trial.

Summary of findings 5. FXIII: Prophylactic trials: intervention compared to inactive control for the prevention and treatment of bleeding in participants without haemophilia

Ir a la tabla de la revisión

Table 1. Red blood cell transfusion requirement

Trial	Number included in analyses intervention / comparator	Clinical condition	Transfusion Protocol?	When was transfusion requirement measured	How were findings reported	Findings: Intervention/ Comparator
Fibrinogen concentrate: prophylactic trials with inactive comparator
Cui 2010	17/14	Surgery (E) : cardiac	Not stated	During ICU stay	Median (IQR) number of red blood cell units transfused	0 (0 ‐ 0)/0 (0 ‐ 1.25)
					Median (IQR) total red blood cell usage	0 (0 ‐ 1.5)/1.5 (0 ‐ 2.3)
Fenger Erikson 2009	10/10	Surgery (E): radical cystectomy	Not stated	During surgery	Median (IQR) number of red blood cell units transfused	2 (0 ‐ 5) / 2.5 (0 ‐ 6)
				48 hours post‐operatively		0 (0 ‐ 2)/1.5 (0 ‐ 2)
				Total transfused		3.5 (0 ‐ 5)/4.0 (0 ‐ 6)
				Within 48 hours post‐operatively	Total number (%) patients requiring a red blood cell transfusion	2 (20%)/8 (80%)
Karlsson 2011	10/10	Surgery (E) : cardiac	Yes	Within 12 hours post‐operatively	Mean (SD) number of red blood cell units transfused	0.2 (0.13)/0.7 (0.27)
Najafi 2014	15/15	Surgery (E): total hip replacement arthroplasty	Yes	"Perioperative period"	Mean (SD) "sum of transfused packed cells"	0.8 (1.01)/1.06 (1.2)
Pournajafian 2015	21/20	Surgery (E): posterior spinal fusion	Not stated	Up to 24 hours post‐operatively	Number (%) patients transfused	0 (0%)/6 (30%)
Ranucci 2015	58/58	Surgery (E): cardiac	Yes	Duration of hospital stay until 30 days	Total number (%) of patients receiving a red blood cell transfusion	19 (33%)/32 (55%)
					Median (IQR) number of units transfused	0 (0 ‐ 1)/1 (0 ‐ 2)
				Within 24 hours of surgery	Mean (SD) mL of red blood cells infused for transfused patients only	136 (178)/356 (488)
				From 24 to 48 hours after surgery	Mean (SD) mL of red blood cells infused for transfused patients only	420 (760)/218 (552)
Sabate 2016	48/44	Surgery (E): liver transplantation	Yes	During surgery	Median (IQR) number of units transfused	0 (0 ‐ 1)/0 (0 ‐ 0.5)
					Total number (%) of patients receiving a red blood cell transfusion	17 (52.9%)/11 (42.7%)
				During and up to 24 hours after surgery	Median (IQR) number of units transfused	2 (0 ‐ 6)/3 (0 ‐ 6)
					Total number (%) of patients receiving a red blood cell transfusion	33 (68.8%)/30 (68.2%)
Sadeghi 2014	30/30	Surgery (E) : cardiac	Yes	Within first 24 hours	Mean (SD) number of red blood cell infusions	1.5 (1.8)/2 (1.5)
					Total number (%) of patients receiving a red blood cell transfusion	15 (50)/21 (70)
Soleimani 2016	31/29	Surgery (E): transurethral resection of the prostate.	Not stated	post‐operatively	Total number (%) of patients receiving a red blood cell transfusion	1 (3.2%)/5 (17.2%)
Fibrinogen concentrate: prophylactic trials with active comparator
Haas 2015	10/9	Surgery (E): scoliosis	Yes	Over 24 hours from the start of surgery	Median (IQR) mL/kg of red blood cells transfused	0 (0 ‐ 15.3)/18.4 (0 ‐ 23.8)
					Mean (SD) mL/kg of red blood cells transfused	3.7 (6.2)/7.2 (8.0)
					Total number (%) of patients receiving a red blood cell transfusion	4 (40%)/6 (67%)
	17/ 13	Surgery (E): craniosynostosis	Yes	Over 24 hours from the start of surgery	Median (IQR) mL/kg of red blood cells transfused	28.2 (21.2‐ 49.9)/55.5 (27.5 ‐ 61.8)
					Mean (SD) mL/kg of red blood cells transfused	33.1 (17.9)/41.8 (18.1)
					Total number (%) of patients receiving a red blood cell transfusion	17 (100%)/13 (100%)
Fibrinogen concentrate: therapeutic trials with inactive comparator
Bilecen 2017	58/59	Surgery (E): cardiac	Yes	Between study intervention and chest closure	Total number (%) of patients receiving a red blood cell transfusion	0 (0%)/3 (5%)
				Between study intervention and 24 hours	Total number (%) of patients receiving a red blood cell transfusion	10 (17%)/20 (33%)
					Median (IQR) number of units transfused	0 (0 ‐ 1)/0 (0 ‐ 4)
Collins 2017	28/27	postpartum haemorrhage	Not stated	From study intervention to discharge	Median (IQR) number of units transfused	1 (0 – 2)/1 (0 – 2)
					Total number (%) of patients who received a red blood cell transfusion	13 (46.6%)/13 (48.1%)
					Total number of red blood cell transfusions	37/38
Curry 2018	24/24	Trauma	Yes	3hrs following admission	Median (IQR) number of units transfused	4 (2 ‐ 6)/2 (2 ‐ 6)
					Total number (%) of patients receiving a red blood cell transfusion	21 (91.3%)/20 (87.0%)
				6hrs following admission	Median (IQR) number of units transfused	3 (2 ‐ 6)/2 (2 ‐ 6)
					Total number (%) of patients receiving a red blood cell transfusion	19 (90.5%)/19 (86.4%)
				24hrs following admission	Median (IQR) number of units transfused	4 (2 ‐ 8)/2 (2 ‐ 5)
					Total number (%) of patients receiving a red blood cell transfusion	19 (95.0%)/18 (85.7%)
Jeppsson 2016	24/24	Surgery (E): cardiac	Yes	During hospital stay	Mean (SD) units of red blood cell s transfused	0.63 (1.17)/1.33 (3.11)
					Median (IQR) number of units transfused	0 (0 ‐ 1)/0 (0 ‐ 1)
Nascimento 2016	21/24	Trauma	Yes	At 24 hours	Median (IQR) number of units transfused	3 (2 ‐ 5)/3 (2 ‐ 4)
Rahe‐Meyer 2013	29/32	Surgery (E): aortic valve replacement	Yes	24 hours after receipt of study medication	Total number (%) of patients receiving a red blood cell transfusion	16 (55)/32 (100)
					Median (IQR) number of units transfused	0 (0 ‐ 3)/2 (2‐5)
Rahe‐Meyer 2016	78/74	Surgery (E): complex cardiac	Yes	24 hours after receipt of study medication	Median (IQR) number of units transfused	1 (0 ‐ 3)/0 (0 ‐ 2)
Wikkelso 2015	123/121	postpartum haemorrhage	Yes	Need for a red blood cell transfusion up to 4 hours post study drug	Total number (%) of patients receiving a red blood cell transfusion	4 (3)/10 (8)
				Need for a red blood cell transfusion up to 24 hours post study drug		14 (11)/19 (16)
				Need for a red blood cell transfusion up to 7 days post study drug		25 (20)/26 (21)
				"During the 6 week period postpartum"		25 (20)/26 (21)
				Time‐point not stated	Total number (%) of patients who received a transfusion of > 4 units of red blood cells.	8 (7)/3 (3)
Fibrinogen concentrate: therapeutic trials with active comparator
Galas 2014	30/33	Surgery (E): cardiac	Yes	post‐operative day 7	Total number (%) of patients receiving a red blood cell transfusion	25 ( 83)/32 (97)
Tanaka 2014	10/10	Surgery (E): valve replacement or repair	Yes	At 24 hours post‐operatively	Total number (%) of patients receiving a red blood cell transfusion	9 (90)/9 (90)
Lance 2012	22/21	Surgery (E): cardiovascular, abdominal or spinal column	Yes	Following study intervention	Mean (SD) number of red blood cell units transfused	4.98 (3.01)/5.38 (2.38)
Innerhofer 2017	50/44	Trauma	Yes	At 24 hours after injury	Total number (%) of patients receiving a red blood cell transfusion	25 (90%)/39 (89%)
					Median (IQR) number of red blood cell units transfused	4 (2 ‐ 7)/6 (4 ‐ 11)
Factor XIII (FXIII): prophylactic trials with inactive comparator
Godje 2006	25/25	Surgery (E) : cardiac	Yes	At 24 hours post‐operatively	Mean (SD) number of red blood cell units transfused	2.6 (1.875)/3.3 (1.397)
Karkouti 2013	138/128	Surgery (E) : cardiac	Not stated	Until post‐operative day 7 or hospital discharge (whichever occurred first)	Median (IQR) number of red blood cell units transfused	0 (0‐1)/0 (0‐1)
					Total number (%) of patients receiving a red blood cell transfusion	46 (33)/43 (34)
Korte 2009	22 in total	Surgery (E): gastrointestinal cancer	Yes	Unclear time point	Median (IQR) mL of red blood cells transfused	150 (0‐700)/200 (0‐800)
Levy 2009	8/8	Surgery (E) : cardiac	Not stated	Until post‐operative day 7 or hospital discharge (whichever occurred first)	Total number of patients requiring allogeneic blood products	No data was reported
					Percentage of patients avoiding a red blood cell transfusion	No data was reported
Rasche 1982	29/31	Acute leukaemia	Yes	Not stated	Median (IQR) number of red blood cell units transfused	8 (0 ‐ 34)/6 (0 ‐ 18)
Factor XIII (FXIII): therapeutic trials with inactive comparator
PCC: prophylactic trials with inactive comparator
(E) = elective surgery; ICU = intensive care unit, IQR = inter quartile range; SD = standard deviation.

Table 1. Red blood cell transfusion requirement

Ir a la tabla de la revisión

Table 2. Blood loss

Trial	Number included in analyses intervention/comparator	Clinical condition	Transfusion Protocol?	How the outcome was measured	Findings: intervention	Findings: Comparator
Fibrinogen concentrate: prophylactic trials with inactive comparator
Cui 2010	17/14	Surgery (E) : cardiac	Yes	Mean (SD) quantity (mL/kg/h) of chest drainage at 1 h post‐operatively	2.9 (2.0) mL/kg/h	3.5 (1.6) mL/kg/h
				Mean (SD) quantity (ml/kg/h) of chest drainage at 6 h post‐operatively	1.3 (0.1) mL/kg/h	1.5 (0.6) mL/kg/h
				Median (IQR) quantity ( mL/kg/h) of chest drainage at 24 h post‐operatively	0.6 (0.4 to 0.8) mL/kg/h	0.7 (0.6 to 0.9) mL/kg/h
Fenger Erikson 2009	10/10	Surgery (E): radical cystectomy	Not stated	Mean (SD) quantity (mL) blood loss at time of intervention	2682 (962) mL	2933 (1320) mL
Karlsson 2011	10/10	Surgery (E) : cardiac	Yes	Mean (SD): "post‐operative bleeding"	565 (150) mL/12 h	830 (268) mL/12 h
Najafi 2014	15/15	Surgery (E): total hip replacement arthroplasty	Yes	Mean (SD) quantity (mL) of intraoperative blood loss	744.6 (337) mL	870 (498) mL
Najafi 2014	15/15	Surgery (E): total hip replacement arthroplasty	Yes	Mean (SD) quantity (mL) of post‐ operative bleeding to 24 hours post surgery	231.6 (67) mL	236.6 (81) mL
Pournajafian 2015	21/20	Surgery (E): posterior spinal fusion	Not stated	Mean (SD) mL blood loss up to 24 hours post‐operatively	533.3 (157.9) mL	679 (130) mL
Ranucci 2015	58/58	Surgery (E): cardiac	Yes	Median (IQR) quantity (mL) lost from chest drain in first 12 hours	300 (200 to 400) mL/12 h	355 (250 to 600) mL/12 h
Sadeghi 2014	30/30	Surgery (E) : cardiac	Yes	Mean (SD) quantity (mL) of blood lost in chest drain in first 12 h post‐operatively	477 (143) mL	703 (179) mL
Soleimani 2016	31/29	Surgery (E): transurethral resection of the prostate.	Not stated	Mean (SD) quantity (mL) of blood lost intraoperatively	521 (290) mL	557 (411) mL
Soleimani 2016	31/29	Surgery (E): transurethral resection of the prostate.	Not stated	Mean (SD) quantity (mL) of blood lost post‐operatively	291 (270) mL	341 (314) mL
Fibrinogen concentrate: prophylactic trials with active comparator
Haas 2015	10/9	Surgery (E): scoliosis	Yes	Median (IQR) "calculated total blood loss (%)" by 24 h post‐operatively	36.5 (14.9 to 54.3) %	51.0 (38.5 to 69.2) %
Haas 2015	17/ 13	Surgery (E): craniosynostosis	Yes	Median (IQR) "calculated total blood loss (%)" by 24 h post‐operatively	89.7 (77.7 to 112.9) %	156.9 (110..5 to 187.3) %
Fibrinogen concentrate: therapeutic trials with inactive comparator
Bilecen 2017	58/59	Surgery (E): cardiac	Yes	Median (IQR) mL blood loss from study intervention to chest closure	50 (29 ‐ 100) mL	70 (33 ‐ 145) mL
				Median (IQR) mL blood loss from ICU arrival to 1hr	70 (35 ‐ 130) mL	90 (46 ‐ 149) mL
				Median (IQR) mL blood loss in ICU from 1hr to 3h	80 (50 ‐ 156) mL	110 (40 ‐ 220) mL
				Median (IQR) mL blood loss in ICU from 3h to 6h	100 (54 ‐ 169) mL	110 (60 ‐ 208) mL
				Median (IQR) mL blood loss in ICU from 6h to 12h	110 (80 ‐ 160) mL	125 (83 ‐ 224) mL
				Median (IQR) mL blood loss in ICU from 12h to 24h	130 (80 ‐ 180) mL	160 (90 ‐ 270) mL
				Median (IQR) cumulative mL blood loss in 24 hours	570 (390 ‐ 730) mL	690 (400 ‐ 1090) mL
Collins 2017	28/27	postpartum haemorrhage	Not stated	Median (IQR) mL blood loss within 24h of study intervention	225 (100 – 341) mL	300 (60 – 800) mL
Jeppsson 2016	24/24	Surgery (E): cardiac	Yes	Mean (SD) mL intraoperative blood loss	306 (110) mL	375 (202) mL
				Median (IQR) mL intraoperative blood loss	300 (200–400) mL	300 (200–500) mL
				Mean (SD) mL post‐operative blood loss (within 12h)	796 (523) mL	897 (553) mL
				Median (IQR) mL post‐operative blood loss (within 12h)	650 (500–835) mL	730 (543–980) mL
				Mean (SD) mL total blood loss (within 12h)	1103 (518) mL	1272 (588) mL
				Median (IQR) mL total blood loss (within 12h)	913 (815–1230) mL	1185 (930–1398) mL
Rahe‐Meyer 2016	78/74	Surgery (E): complex cardiac	Yes	Median (IQR) mL chest tube drainage volume at 6h	260.0 (155.0–410.0) mL	297.5 (200.0–455.0) mL
				Median (IQR) mL chest tube drainage volume at 12h	405.0 (245.0–600.0) mL	447.5 (320.0–700.0) mL
				Median (IQR) mL chest tube drainage volume at 24h	590.0 (400.5–839.5) mL	682.5 (530.0–1050.0) mL
Wikkelso 2015	123/121	postpartum haemorrhage	Yes	Median (IQR) quantity (mL) lost to 6 weeks postpartum	1700 (1500 to 2000) mL	1700 (1400 to 2000) mL
Fibrinogen concentrate: therapeutic trials with active comparator
Galas 2014	30/33	Surgery (E): cardiac	Yes	Median (IQR) quantity (mL) chest drainage from intraoperative to 48 h post‐operatively	320 (157 ‐ 750) mL	410 (215 ‐ 510) mL
Tanaka 2014	10/10	Surgery (E): valve replacement or repair	Yes	Median (IQR) quantity (mL) chest tube drainage at 12 h post‐operatively	925 (500 to 1693) mL	1315 (653 to 2965) mL
Factor XIII (FXIII): prophylactic trials with inactive comparator
Godje 2006	25/25	Surgery (E) : cardiac	Yes	Mean (SD) drain volume (ml) at 6 h post‐operatively^$	231 (113) mL	278 (98) mL
				Mean (SD) drain volume (ml) at 12 h post‐operatively^$	371 (170) mL	458 (160) mL
				Mean (SD) drain volume (ml) at 24 h post‐operatively^$	649 (227) mL	747 (222) mL
				Mean (SD) drain volume (ml) at 36 h post‐operatively^$	798 (273) mL	969 (309) mL
				Mean (SD) drain volume (ml) at 48h post‐operatively^$	958 (376) mL	1231 (175) mL
Korte 2009	22 in total	Surgery (E): gastrointestinal cancer	Yes	Median (IQR) quantity (ml) blood loss at completion of surgery	750 (400 to 1000) mL	1050 (700 to 1800) mL
Levy 2009	8/8	Surgery (E) : cardiac	Not stated	Mean (SD) quantity (mL) of chest drainage at 8 h post‐operatively	358 (93.5) mL	505 (301) mL
				Mean (SD) quantity (mL) of chest drainage at 24h post‐operatively	903 (738) mL	1155 (804) mL
				Mean (SD) quantity (mL) of chest drainage at CT removal	1007 (912) mL	1489 (1395) mL
Factor XIII (FXIII): therapeutic trials with inactive comparator
PCC: prophylactic trials with inactive comparator
^ "Calculated total blood loss (%) was based on the estimated total blood volume (calculated as described by Kearney 1989 ) " Haas 2015 * "The visual assessment of the surgical field was performed by the senior surgical staff as follows: 0 = excellent haemostasis (dry field), 1 = mild bleeding (oozing), 2 = moderate bleeding (controllable with applied pressure) and 3 = severe bleeding (multiple diffuse bleeding sites)" Tanaka 2014 $ Findings are reported graphically, not numerically. (E) = elective surgery; h = hours; ICU = intensive care unit, IQR = inter quartile range; mL = millilitres; SD = standard deviation.

Table 2. Blood loss

Ir a la tabla de la revisión

Table 3. Fibrinogen concentrate: prophylactic trials with inactive comparator

Trial	Number randomised (N = intervention/N = comparator)	Number included in analyses (N = intervention/N = comparator)	Intervention	Comparator	Co‐interventions	Primary Outcomes	Clinical Condition	Population Age
Cui 2010	40 (20/20)	31 (17/14)	Fibrinogen concentrate (0.5‐1g) combined with traditional transfusion, guided by TEG	NONE: Traditional transfusion guided by clinical experience.	Protamine, FFP, red blood cells, platelets.	Time to chest wall closure.	Surgery (E) : cardiac	Paediatric population
Fenger Erikson 2009	21 (10/11 )	20 (10/10)	Fibrinogen concentrate (45mg/kg), administered intra‐operatively	PLACEBO: isotonic saline, administered intra‐operatively at a dose of 2.25 mL/kg).	Intra‐operative FFP, red blood cells.	Whole blood maximum clot firmness as determined by thromboelastometry.	Surgery (E): radical Cystectomy	Adult population
Karlsson 2011	20 (10/10)	20 (10/10)	Preoperative infusion offibrinogen concentrate (2g) at 5 minutes following baseline measurements.	NONE	Heparin, protamine, red blood cells, aspirin.	Clinical adverse events, including graft occlusion on CT up to 3 to 4d post‐ operatively.	Surgery (E) : cardiac	Adult population
Najafi 2014	30 (15/15)	30 (15/15)	Fibrinogen concentrate (30 mg/kg) infused after induction of general anaesthesia	PLACEBO: normal saline administered in equal volume as fibrinogen.	None stated.	Volume of red blood cells transfused during and 24h post‐operatively.	Surgery (E): total hip replacement arthroplasty	Adult population
Pournajafian 2015	41 (21/20)	41 (21/20)	Fibrinogen concentrate (1g) infused after induction of general anaesthesia.	PLACEBO: normal saline.	None stated.	Not stated.	Surgery (E): posterior spinal fusion.	Adult population
Ranucci 2015	116 (58/58)	116 (58/58	Fibrinogen concentrate with dose determined by the mean clot firmness test on the FIBTEM.	PLACEBO: Normal (0.9%) saline.	TXA during surgery, protamine.	Avoidance of allogeneic blood products transfusion during hospital stay up to 30d.	Surgery (E): cardiac	Adult population
Sabate 2016	99 (51/48)	92 (48/44)	Fibrinogen concentrate with dose calculated to reach a target plasma concentration of 2.9 g/L.	PLACEBO: Normal (0.9%) saline.	Red blood cells, platelets, FFP, tranexamic acid.	The percentage of patients requiring transfusion of red blood cell units during the liver transplant procedure.	Surgery (E): liver transplantation.	Adult population
Sadeghi 2014	60 (30/30)	60 (30/30)	Fibrinogen (1g) administered 30 minutes prior to anaesthesia induction.	PLACEBO: saline administered 30 minutes prior to anaesthesia induction	None stated.	Volume of post‐operative haemorrhage 0, 12 and 24h post‐operatively.	Surgery (E): cardiac	Adult population
Soleimani 2016	72 (36/36)	60 (31/29)	Fibrinogen concentrate (2g) administered 15‐30 minutes before the start of surgery.	PLACEBO: 50ml normal saline.	None stated.	Bleeding volume during and after surgery.	Surgery (E): transurethral resection of the prostate.	Adult population
* For the outcomes bleeding and transfusion requirements, 1 patient from the Factor XIII, 1250 IE arm and 2 patients from the placebo arm were excluded from the analysis due to re‐operations following hints of early bypass occlusions. All patients were included in the analysis of the other outcomes. d = days; (E) = elective surgery; FIBTEM = is a ready‐to‐use ROTEM^® system reagent for use with citrated whole blood and assesses the clot firmness of the fibrin clot, h = hours; PCC = prothrombin complex concentrate; mins = minutes; wks = weeks.

Table 3. Fibrinogen concentrate: prophylactic trials with inactive comparator

Ir a la tabla de la revisión

Table 4. Fibrinogen concentrate: prophylactic trials with active comparator

Trial

Number randomised

(N = intervention/ N = comparator)

Number included in analyses

(N = intervention/ N = comparator)

Intervention

Comparator

Co‐interventions

Primary Outcomes

Clinical Condition

Population Age

Haas 2015

Craniosynostosis surgery = 36

(17/14).

Scoliosis surgery = 26

(10/9)

Craniosynostosis surgery = 36

(17/13).

Scoliosis surgery = 26

(10/9)

Fibrinogen concentrate if FIBTEM MCF under 13mm (early substitution group).

Total dose administered per subject in

craniosynostosis surgery; median of 90mg kg^‐1 (IQR, 75 to 120 mg kg^‐1);

Scoliosis surgery = = median of 60mg kg^‐1 (IQR, 30 to 68 mg kg^‐1).

Fibrinogen concentrate if FIBTEM MCF under 8 mm (conventional).

Total dose administered per subject in

craniosynostosis surgery = median of

90mg kg^‐1 (IQR, 60 to 90 mg kg^‐1); Scoliosis surgery = median of 30mg kg^‐1 (IQR, 30 to 60 mg kg^‐1).

Tranexamic acid

Total volume of transfused red blood cells per kg bodyweight within 24h post‐operatively

Surgery (E): craniosynostosis or scoliosis

Paediatric population

Table 4. Fibrinogen concentrate: prophylactic trials with active comparator

Ir a la tabla de la revisión

Table 5. Fibrinogen concentrate: therapeutic trials with inactive comparator

Trial	Number randomised (N = intervention/ N = comparator)	Number included in analyses (N = intervention/ N = comparator)	Intervention	Comparator	Co‐interventions	Primary Outcomes	Clinical Condition	Population Age
Bilecen 2017	120 (60/60)	115 (58/57)	Fibrinogen concentrate, dose calculated to provide a target plasma fibrinogen concentration of 2.5 g/L.	PLACEBO: 2g of albumin diluted with 50ml 0.9% sodium chloride solution.	Red blood cells, platelets, FFP, tranexamic acid.	The intraoperative blood loss measured between intervention and closure of the chest.	Surgery (E): cardiac	Adult population
Collins 2017	57 (29/28)	55 (28/27)	Fibrinogen concentrate, dose determined by FIBTEM to elevate A5 result above 22mm.	PLACEBO: 50ml of normal saline.	Red blood cells, platelets, FFP, tranexamic acid.	The number of allogeneic blood products (red blood cell , FFP, cryoprecipitate, platelets) infused after study medication until hospital discharge.	postpartum haemorrhage	Adult population
Curry 2018	48 (24/24)	48 (24/24)	Fibrinogen concentrate 6g infusion.	PLACEBO: 300ml of normal saline.	Major haemorrhage therapy as per hospital protocol: red blood cells, platelets, FFP, tranexamic acid, cryoprecipitate.	Feasibility of delivery of FgC therapy within 45 minutes to adult trauma patients and the proportion of participants whose fibrinogen levels were maintained ≥ 2 g/L during active haemorrhage.	Trauma	Adult population
Jeppsson 2016	52 (26/26)	48 (24/24)	Fibrinogen (2g) as an infusion over 5 minutes, immediately before surgery.	PLACEBO: 100ml of normal saline.	300 units kg−1 of heparin, 1 mg protamine per 100 units of heparin. Red blood cells, platelets, FFP, additional fibrinogen.	Mediastinal drain loss during the first 12 hours post‐operatively.	Surgery (E): cardiac	Adult population
Nascimento 2016	50 (25/25)	45 (21/24)	Fibrinogen concentrate, 6g as a 3 minute infusion.	PLACEBO: 300ml of normal saline as a 3 minute infusion.	Red blood cells, tranexamic acid, FFP, platelets and cryoprecipitate.	The feasibility of providing the study intervention within 1 hour of hospital admission.	Trauma	Adult population
Rahe‐Meyer 2013	80 (38 /42)	71 (29/32)	Fibrinogen concentrate, dose determined by FIBTEM test: median (IQR range) dose: 8g (6 to 9g).	PLACEBO: normal (0.9%) saline. Median (IQR range) dose: 400ml (300 to 450ml).	TXA during surgery, platelets or FFP.	Total number of units of allogeneic blood components given during first 24h post‐operatively	Surgery (E): aortic valve replacement	Adult population
Rahe‐Meyer 2016	152 (78/74)	152 (78/74)	Fibrinogen concentrate, dose determined by FIBTEM testing to generate an MCF of 22mm.	PLACEBO: 0.9% sodium chloride solution.	Red blood cells, platelets, FFP, tranexamic acid.	The number of units of allogeneic blood products (FFP, platelets and red blood cells) administered during the 24 hours after administration of study medication.	Surgery (E): complex cardiac	Adult population
Wikkelso 2015	249 (124 /125)	249 (PPA: 123 /121) (ITT: 120 /119)	Fixed dose of 2g of fibrinogen concentrate dispensed using a syringe pump infusion over 20 minutes	PLACEBO: 100ml of isotonic saline (mean (standard deviation) dose =1.2 (0.2) mg/kg).	TXA, hydroxyethyl starch, intravenous fluids.	Red blood cell transfusion during a 6 wk follow‐up period postpartum	postpartum haemorrhage	Adult population
d = days; (E) = elective surgery; FFP = fresh frozen plasma; FIBTEM = is a ready‐to‐use ROTEM^® system reagent for use with citrated whole blood and assesses the clot firmness of the fibrin clot; h = hours; INR = International Normalised Ratio, IQR = inter quartile range; mins = minutes; PCC = prothrombin complex concentrate; TXA = tranexamic acid; (U) = urgent surgery/ treatment; wks = weeks.

Table 5. Fibrinogen concentrate: therapeutic trials with inactive comparator

Ir a la tabla de la revisión

Table 6. Fibrinogen concentrate: therapeutic trials with active comparator

Trial	Number randomised (N = intervention/ N = comparator)	Number included in analyses (N = intervention/ N = comparator)	Intervention	Comparator	Co‐interventions	Primary Outcomes	Clinical Condition	Population Age
Galas 2014	63 (30/33)	63 (30/33)	Single dose of Fibrinogen concentrate (60mg/kg body weight, Haemocomplettan®) given at time of intra‐operative bleeding.	Single dose of Cryoprecipitate (10ml/kg body weight) given at time of intra‐operative bleeding.	Not Stated	post‐operative blood losses during 48h after surgery	Surgery (E): cardiac	Paediatric population
Lance 2012	43 (22/21)	43 (22/21) (but only 16/16 for primary outcome analysis).	FFP ( 2 units) + Fibrinogen (2g)	FFP (4 units)	Protamine, red blood cells	Pre‐ and post‐transfusion ROTEM analysis	Surgery (E): cardiovascular, abdominal or spinal column	Adult population
Tanaka 2014	20 (10/10)	20 (10/10)	A single dose (4g) of Fibrinogen (RiaSTAP®) given within 30 minutes of intervention decision.	One unit of apheresis platelets (median volume 230 ml) within 30 minutes of intervention decision.	Red blood cells, platelets, plasma, cryoprecipitate and fibrinogen if triggers met	i) Haemostatic condition in the surgical field post‐intervention; ii) Haemostatic blood product use over the first 24h; iii) Percentage of patients with thromboembolic events at 6‐8 weeks post‐operatively; iv) Mortality at 6‐8 weeks post‐operatively	Surgery (E): valve replacement or repair	Adult population
Innerhofer 2017	100 (52/48)	94 (50/44)	Fibrinogen concentrate (50 mg/kg of bodyweight) for abnormal ROTEM fibrin polymerisation (FibA10 < 9 mm) or four‐Factor PCC (20 IU/kg of bodyweight) for delayed initial thrombin formation (ExCT > 90s or prothrombin time index < 35%). FXIII concentrate (20 IU/kg of bodyweight) was administered with each second fibrinogen dose.	FFP (15 ml/kg bodyweight).	Tranexamic acid 20mg/kg, FXIII concentrate (20 IU/kg of bodyweight),	Occurrence of multiple organ failure during ICU stay as assessed by the daily SOFA score.	Trauma	Adult population
* 20 of the 47 patients had a 2nd dose of 20ml of PCC: in 3 because of continued bleeding and in 17 due to INR target not being met. 3 of the 47 patients received a 3rd dose of PCC due to an unmet INR level. ^ 3 of the 46 patients had a 2nd dose of 20 ml PCC: 1 due to a continued bleed and 2 did not reach the target INR. d = days; (E) = elective surgery; FFP = fresh frozen plasma; h = hours; INR = International Normalised Ratio; mins = minutes; PCC = prothrombin complex concentrate; ROTEM^® = rotational thromboelastometry; TXA = tranexamic acid; (U) = urgent surgery/ treatment; wks = weeks.

Table 6. Fibrinogen concentrate: therapeutic trials with active comparator

Ir a la tabla de la revisión

Table 7. Factor XIII (FXIII): prophylactic trials with inactive comparator

Trial	Number randomised (N = intervention/ N = comparator)	Number included in analyses (N = intervention/ N = comparator)	Intervention	Comparator	Co‐interventions	Primary Outcomes	Clinical Condition	Population Age
Godje 2006	75 (25/25/ 25)	75 (25/25/25)*	Infusion of FactorXIII1) 1250 units or2) 2500 unitsimmediately after administration of protamine (protamine was administered during the scheduled coronary artery bypass graft surgery) This is the group that was included in this review.	NONE: standard surgical and post‐operative care	Protamine for heparin reversal FFP, red blood cells, platelets	Not stated	Surgery (E) : cardiac	Adult population
Karkouti 2013	479 (164/162/153)	409 (143/138/128)	A single intravenous dose of FactorXIII 1) 17.5 IU/kg or2) 35 IU/kgbefore the induction of anaesthesia. This is the group that was included in this review.	PLACEBO: single dose of an unspecified placebo before the induction of anaesthesia.	Protamine, antifibrinolytics during cardiopulmonary bypass, blood products	Percentage of patients avoiding any allogeneic transfusions within 7d or discharge home.	Surgery (E) : cardiac	Adult population
Korte 2009	25 (no further details stated)	22 (no further details stated)	FactorXIII (30 U/kg) infused 15 minutes after the beginning of surgery.	PLACEBO: albumin infusion.	Low molecular weight heparin, FFP, red blood cells, platelets	Clot firmness over 195 mins of surgery	Surgery (E): gastrointestinal cancer	Adult population
Levy 2009	43 (10/9/8/8)	43 (10/9/8/8)	A single infusion of FactorXIII at 4 doses:1) 11.9 units/kg;2) 25 units/kg;3) 35 units/kg;4) 50 units/kg. This is the group that was included in this review.	PLACEBO: no further details given.	Protamine for heparin reversal FFP, red blood cells, platelets Cryoprecipitate, Fibrinogen	Incidence and severity of adverse events from intervention administration to follow‐up at 5 to 7 wks	Surgery (E) : cardiac	Adult population
Rasche 1982	60 (29/31)	60 (29/31)	FactorXIII (1000 units) on day 1 of treatment and a subsequent daily dose of 500 units plus normal care protocols.	NONE: normal care products	Red blood cells, platelets, whole blood	Bleeding complications	Acute leukaemia	Adult population
Shirahata 1990	58 (30/28)	58 (30/28)	FactorXIII (dose of 70 to 100 units) administered slowly for at least 10 minutes intravenously within 6 hours of delivery.	NONE	Not stated	Not stated	Prematurity	Infant population

Table 7. Factor XIII (FXIII): prophylactic trials with inactive comparator

Ir a la tabla de la revisión

Table 8. Factor XIII (FXIII): therapeutic trials with inactive comparator

Trial

Number randomised (N = intervention/ N = comparator)

Number included in analyses (N = intervention/ N = comparator)

Intervention

Comparator

Co‐interventions

Primary Outcomes

Clinical Condition

Population Age

Bregenzer 1999

(17/11)

(11/9)

FactorXIII as an i.v. injection (at 3750 units on Day 0 and 1250 units on days 1 through 9) as an addition to basic steroid treatment.

PLACEBO: continuation of basic steroid treatment

TXA,

platelets or FFP post‐intervention

Time to cessation of macroscopically visible intestinal bleeding within 14d of starting treatment.

Ulceratice colitis

Adult population

Table 8. Factor XIII (FXIII): therapeutic trials with inactive comparator

Ir a la tabla de la revisión

Table 9. PCC: prophylactic trials with inactive comparator

Trial

Number randomised (N = intervention/ N = comparator)

Number included in analyses (N = intervention/ N = comparator)

Intervention

Comparator

Co‐interventions

Primary Outcomes

Clinical Condition

Population Age

Turner 1981

(39/39)

PCC (dose of 1ml/kg or 2ml/kg)

NONE: normal care protocols.

Vitamin K, heparin and clotting Factors for disseminated intravascular coagulation, red blood cells

i) Mortality (all‐cause)

ii) Incidence of intraventricular haemorrhage (post‐mortem)

Prematurity

Infant population

Table 9. PCC: prophylactic trials with inactive comparator

Ir a la tabla de la revisión

Comparison 1. Fibrinogen: prophylactic trials: intervention vs inactive control

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 All‐cause mortality Show forest plot	4		Risk Ratio (M‐H, Fixed, 95% CI)	Totals not selected

1.1 Up to 4 days post‐operatively	1		Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
1.2 Up to 7 days post‐operatively	1		Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
1.3 Up to 30 days post‐operatively	1		Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
1.4 Prior to hospital discharge	1		Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
2 Arterial thromboembolic events Show forest plot	7		Risk Ratio (M‐H, Fixed, 95% CI)	Totals not selected

2.1 Up to 3 days post‐operatively	1		Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
2.2 Up to 4 days post‐operatively	2		Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
2.3 Up to 7 days post‐operatively	1		Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
2.4 Up to 30 days post‐operatively in patients who have had cardiac surgery	1		Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
2.5 Up to 30 days post‐operatively in patients who had undergone elective surgery	2		Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
3 Venous thromboembolic events Show forest plot	7		Risk Ratio (M‐H, Fixed, 95% CI)	Totals not selected

3.1 Up to 3 days post‐operatively	1		Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
3.2 Up to day 4 post‐operatively	2		Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
3.3 Up to 7 days post‐operatively	1		Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
3.4 Up to 30 days post‐operatively in patients who have had cardiac surgery	1		Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
3.5 Up to 30 days post‐operatively in participants who had undergone elective surgery	2		Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
4 Mortality due to bleeding Show forest plot	4		Risk Ratio (M‐H, Fixed, 95% CI)	Totals not selected

4.1 Up to 3 days post‐operatively	1		Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
4.2 Up to 7 days post‐operatively	1		Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
4.3 Up to 30 days post‐operatively	1		Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
4.4 Prior to hospital discharge	1		Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
5 RBC transfusion requirement Show forest plot	3		Mean Difference (IV, Fixed, 95% CI)	Totals not selected

5.1 Intra‐operatively	1		Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
5.2 Up to 12 hours post‐operatively	1		Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
5.3 Up to 24 hours post‐operatively	1		Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
5.4 In peri‐operative period (unclear time‐point)	1		Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
6 Receipt of a RBC transfusion Show forest plot	8		Risk Ratio (M‐H, Fixed, 95% CI)	Subtotals only

6.1 Intra‐operatively	1	92	Risk Ratio (M‐H, Fixed, 95% CI)	1.42 [0.75, 2.68]
6.2 Up to 12 hours post‐operatively	1	20	Risk Ratio (M‐H, Fixed, 95% CI)	0.33 [0.04, 2.69]
6.3 Up to 24 hours post‐operatively	3	207	Risk Ratio (M‐H, Fixed, 95% CI)	0.53 [0.37, 0.76]
6.4 Up to 24 hours post‐operatively	1	41	Risk Ratio (M‐H, Fixed, 95% CI)	0.07 [0.00, 1.22]
6.5 During surgery and up to 24 hours post‐operatively	1	92	Risk Ratio (M‐H, Fixed, 95% CI)	1.01 [0.76, 1.33]
6.6 Up to 48 hours post‐operatively in patients who have had cardiac surgery	1	116	Risk Ratio (M‐H, Fixed, 95% CI)	0.59 [0.38, 0.92]
6.7 Up to 48 hours post‐operatively in patients who have undergone elective, non‐cardiac surgery	2	80	Risk Ratio (M‐H, Fixed, 95% CI)	0.23 [0.07, 0.68]
7 Blood loss Show forest plot	7		Ratio of Geometric Means (Fixed, 95% CI)	Subtotals only

7.1 Intraoperative blood loss (mL)	2		Ratio of Geometric Means (Fixed, 95% CI)	0.96 [0.77, 1.21]
7.2 Chest tube drainage at 1 hour post operatively (ml/kg/hr)	1		Ratio of Geometric Means (Fixed, 95% CI)	0.75 [0.52, 1.09]
7.3 Chest tube drainage at 6 hours post‐operatively (ml/kg/hr)	1		Ratio of Geometric Means (Fixed, 95% CI)	0.93 [0.76, 1.14]
7.4 Chest tube drainage in first 12 hours post‐operatively (mL)	2		Ratio of Geometric Means (Fixed, 95% CI)	0.68 [0.60, 0.76]
7.5 Bleeding up to 24 hours post‐operatively (mL)	2		Ratio of Geometric Means (Fixed, 95% CI)	0.83 [0.73, 0.94]
7.6 Bleeding at up to 48 hours post‐operatively (mL)	1		Ratio of Geometric Means (Fixed, 95% CI)	0.85 [0.57, 1.27]
7.7 Bleesing during surgery and up to 48 hours post‐operatively (mL)	1		Ratio of Geometric Means (Fixed, 95% CI)	0.94 [0.67, 1.33]
8 Allergic Adverse Events Show forest plot	3		Odds Ratio (M‐H, Fixed, 95% CI)	Totals not selected

8.1 Up to 2 days post‐operatively	1		Odds Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
8.2 Up to 3 days post‐operatively	1		Odds Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
8.3 Up to 30 days post‐operatively	1		Odds Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]

Comparison 1. Fibrinogen: prophylactic trials: intervention vs inactive control

Ir a la tabla de la revisión

Comparison 2. Fibrinogen: prophylactic trials: intervention vs haemostatically active control

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 All‐cause mortality Show forest plot	1		Odds Ratio (M‐H, Fixed, 95% CI)	Totals not selected

1.1 Duration of hospital stay (whole population)	1		Odds Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
2 Arterial thromboembolic events Show forest plot	1		Odds Ratio (M‐H, Fixed, 95% CI)	Totals not selected

2.1 Duration of hospital stay (whole population)	1		Odds Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
3 Venous thromboembolic events Show forest plot	1		Odds Ratio (M‐H, Fixed, 95% CI)	Totals not selected

3.1 Duration of hospital stay (whole population)	1		Odds Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
4 Mortality due to bleeding Show forest plot	1		Odds Ratio (M‐H, Fixed, 95% CI)	Totals not selected

4.1 Duration of hospital stay (whole population)	1		Odds Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
5 RBC transfusion requirement Show forest plot	1		Mean Difference (IV, Fixed, 95% CI)	Totals not selected

5.1 Surgical time plus 24 hours post‐operatively ‐ Craniosynostosis (ml/kg)	1		Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
5.2 Surgical time plus 24 hours post‐operatively ‐ Scoliosis (ml/kg)	1		Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
6 Receipt of a RBC transfusion Show forest plot	1		Risk Ratio (M‐H, Fixed, 95% CI)	Totals not selected

6.1 Surgical time plus 24 hours post‐operatively ‐ Craniosynostosis	1		Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
6.2 Surgical time plus 24 hours post‐operatively ‐ Scoliosis	1		Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
7 Allergic Adverse Events Show forest plot	1		Odds Ratio (M‐H, Fixed, 95% CI)	Totals not selected

7.1 Up to 24 hours post‐operatively (whole population)	1		Odds Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]

Comparison 2. Fibrinogen: prophylactic trials: intervention vs haemostatically active control

Ir a la tabla de la revisión

Comparison 3. Fibrinogen: therapeutic trials: intervention vs inactive control

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 All‐cause mortality Show forest plot	7		Risk Ratio (M‐H, Fixed, 95% CI)	Subtotals only

1.1 Up to 28 days following admission	2	97	Risk Ratio (M‐H, Fixed, 95% CI)	1.46 [0.71, 2.99]
1.2 In‐hospital mortality up to 30 days post‐operatively	1	120	Risk Ratio (M‐H, Fixed, 95% CI)	5.0 [0.25, 102.00]
1.3 Up to 6 weeks postnatally	2	294	Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
1.4 Up to 46 days post‐operatively	2	213	Risk Ratio (M‐H, Fixed, 95% CI)	0.23 [0.05, 1.01]
2 Arterial thromboembolic events Show forest plot	7		Risk Ratio (M‐H, Fixed, 95% CI)	Totals not selected

2.1 During hospital stay	1		Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
2.2 Up to 28 days following admission	2		Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
2.3 Up to 30 days post‐operatively	1		Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
2.4 Up to 45 days post‐operatively	1		Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
2.5 Up to 6 weeks postnatally	2		Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
3 Venous thromboembolic events Show forest plot	6		Risk Ratio (M‐H, Fixed, 95% CI)	Totals not selected

3.1 During hospital stay	1		Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
3.2 Up to 28 days following admission	1		Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
3.3 Up to 30 days post‐operatively	1		Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
3.4 Up to 45 days post‐operatively	1		Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
3.5 Up to 6 weeks postnatally	2		Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
4 Mortality due to bleeding Show forest plot	5		Risk Ratio (M‐H, Fixed, 95% CI)	Subtotals only

4.1 Up to 28 days following admission	2	93	Risk Ratio (M‐H, Fixed, 95% CI)	2.45 [0.38, 15.76]
4.2 Up to 6 weeks postnatally	2	294	Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
4.3 Up to 46 days post‐operatively	1	152	Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
5 Receipt of a RBC transfusion Show forest plot	6		Risk Ratio (M‐H, Fixed, 95% CI)	Subtotals only

5.1 Intraoperative: Between study medication and chest closure	1	117	Risk Ratio (M‐H, Fixed, 95% CI)	0.15 [0.01, 2.75]
5.2 Up to 3 hours following admission	1	46	Risk Ratio (M‐H, Fixed, 95% CI)	1.05 [0.86, 1.29]
5.3 Within 4 hours of receipt of the study drug	1	244	Risk Ratio (M‐H, Fixed, 95% CI)	0.39 [0.13, 1.22]
5.4 Up to 6 hours following admission	1	43	Risk Ratio (M‐H, Fixed, 95% CI)	1.05 [0.84, 1.30]
5.5 Up to 24 hours following admission	1	41	Risk Ratio (M‐H, Fixed, 95% CI)	1.11 [0.91, 1.36]
5.6 Within 24 hours of receipt of study drug in patients who have had cardiac surgery	2	178	Risk Ratio (M‐H, Fixed, 95% CI)	0.54 [0.39, 0.75]
5.7 Within 24 hours of receipt of study drug in women who have experienced a post‐partum haemorrhage	1	244	Risk Ratio (M‐H, Fixed, 95% CI)	0.72 [0.38, 1.38]
5.8 Within 7 days of receipt of the study drug	1	244	Risk Ratio (M‐H, Fixed, 95% CI)	0.95 [0.58, 1.54]
5.9 From end of intervention to hospital discharge	1	55	Risk Ratio (M‐H, Fixed, 95% CI)	0.96 [0.55, 1.69]
5.10 During hospital stay	1	48	Risk Ratio (M‐H, Fixed, 95% CI)	1.0 [0.38, 2.66]
6 RBC transfusion requirement Show forest plot	1		Mean Difference (IV, Fixed, 95% CI)	Totals not selected

6.1 During hospital stay	1		Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
7 Blood loss Show forest plot	1		Ratio of Geometric Mean (Fixed, 95% CI)	Totals not selected

7.1 Intra‐operative	1		Ratio of Geometric Mean (Fixed, 95% CI)	0.0 [0.0, 0.0]
7.2 Post‐operative (within 12 hours)	1		Ratio of Geometric Mean (Fixed, 95% CI)	0.0 [0.0, 0.0]
7.3 Intra‐ and post‐operative (within 12 hours)	1		Ratio of Geometric Mean (Fixed, 95% CI)	0.0 [0.0, 0.0]
8 Allergic Adverse Events Show forest plot	4		Risk Difference (M‐H, Fixed, 95% CI)	Totals not selected

8.1 Up to 24 hours from study medication	1		Risk Difference (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
8.2 Up to 10 days from study medication	1		Risk Difference (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
8.3 Up to 28 days following admission	1		Risk Difference (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
8.4 Up to 30 days from study medication	1		Risk Difference (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]

Comparison 3. Fibrinogen: therapeutic trials: intervention vs inactive control

Ir a la tabla de la revisión

Comparison 4. Fibrinogen: therapeutic trials: intervention vs haemostatically active control

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 All‐cause mortality Show forest plot	4		Risk Ratio (M‐H, Fixed, 95% CI)	Totals not selected

1.1 Up to 7 days post‐operatively or hospital discharge	1		Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
1.2 Up to 28 days post‐operatively	1		Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
1.3 Up to 30 days post‐operatively in patients who have undergone major elective surgery	1		Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
1.4 Up to 30 days post‐operatively in trauma patients	1		Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
2 Arterial thromboembolic events Show forest plot	3		Risk Ratio (M‐H, Fixed, 95% CI)	Totals not selected

2.1 Up to 7 days post‐operatively	1		Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
2.2 Up to 28 days post‐operatively	1		Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
2.3 Up to 30 days post‐operatively	1		Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
3 Venous thromboembolic events Show forest plot	4		Odds Ratio (M‐H, Fixed, 95% CI)	Totals not selected

3.1 Up to 7 days post‐operatively	1		Odds Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
3.2 Up to 28 days post‐operatively	1		Odds Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
3.3 Up to 30 days post‐operatively	1		Odds Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
3.4 Up to day 30 following trauma	1		Odds Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
4 Mortality due to bleeding Show forest plot	4		Risk Ratio (M‐H, Fixed, 95% CI)	Subtotals only

4.1 Up to 7 days post‐operatively	1	63	Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
4.2 Up to 30 days post‐operatively	2	63	Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
4.3 Up to day 30 following trauma	1	94	Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
5 RBC transfusion requirement Show forest plot	1		Mean Difference (IV, Fixed, 95% CI)	Totals not selected

5.1 Up to 30 days post‐operatively	1		Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
6 Receipt of a RBC transfusion Show forest plot	3		Risk Ratio (M‐H, Fixed, 95% CI)	Totals not selected

6.1 Up to 24 hours post‐operatively	1		Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
6.2 Up to 24 hours following trauma	1		Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
6.3 Up to 7 days post‐operatively	1		Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
7 Allergic Adverse Events Show forest plot	2		Odds Ratio (M‐H, Fixed, 95% CI)	Totals not selected

7.1 Up to 7 days post‐operatively	1		Odds Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
7.2 Up to 30 days post‐operatively	1		Odds Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]

Comparison 4. Fibrinogen: therapeutic trials: intervention vs haemostatically active control

Ir a la tabla de la revisión

Comparison 5. FXIII: prophylactic trials: intervention vs inactive control

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 All‐cause mortality Show forest plot	5		Risk Ratio (M‐H, Fixed, 95% CI)	Totals not selected

1.1 Until hospital discharge	1		Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
1.2 Up to 5 to 7 weeks post‐operatively	2		Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
1.3 Up to 340 days post‐operatively	1		Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
1.4 Time point not stated	1		Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
2 Arterial thromboembolic events Show forest plot	4		Risk Ratio (M‐H, Fixed, 95% CI)	Totals not selected

2.1 Up to 5 to 7 weeks post‐operatively	2		Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
2.2 Up to 30 days post‐operatively	1		Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
2.3 Post‐operatively	1		Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
3 Venous thromboembolic events Show forest plot	4		Risk Ratio (M‐H, Fixed, 95% CI)	Totals not selected

3.1 Up to 5 to 7 weeks post‐operatively	2		Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
3.2 Up to 30 days post‐operatively	1		Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
3.3 Post‐operatively	1		Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
4 Mortality due to bleeding Show forest plot	4		Risk Ratio (M‐H, Fixed, 95% CI)	Totals not selected

4.1 Up to a median of 32 days from study intervention	1		Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
4.2 Until hospital discharge	1		Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
4.3 Up to 5 to 7 weeks post‐operatively	1		Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
4.4 Up to 340 days post‐operatively	1		Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
5 RBC transfusion requirement Show forest plot	1		Mean Difference (IV, Fixed, 95% CI)	Totals not selected

5.1 From the point of study medication	1		Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
6 Receipt of a RBC transfusion Show forest plot	1		Risk Ratio (M‐H, Fixed, 95% CI)	Totals not selected

6.1 Up to 7 days post‐operatively (or discharge)	1		Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
7 Blood loss Show forest plot	2		Ratio of Geometric Mean (Fixed, 95% CI)	Subtotals only

7.1 Chest tube drainage at 8 hours post‐operatively (mL)	2		Ratio of Geometric Mean (Fixed, 95% CI)	0.79 [0.65, 0.97]
7.2 Chest drain volume at 12 hours post‐operatively (mL)	1		Ratio of Geometric Mean (Fixed, 95% CI)	0.78 [0.63, 0.97]
7.3 Chest tube drainage at 24 hours post‐operatviely (mL)	2		Ratio of Geometric Mean (Fixed, 95% CI)	0.85 [0.72, 1.00]
7.4 Chest drain volume at 36 hours post‐operatively (mL)	1		Ratio of Geometric Mean (Fixed, 95% CI)	0.82 [0.68, 0.98]
7.5 Chest drain volume at 48 hours post‐operatively (mL)	1		Ratio of Geometric Mean (Fixed, 95% CI)	0.73 [0.62, 0.86]
7.6 Chest drainage at drain removal (mL)	1		Ratio of Geometric Mean (Fixed, 95% CI)	0.68 [0.32, 1.47]
8 Allergic Adverse Events Show forest plot	5		Odds Ratio (M‐H, Fixed, 95% CI)	Subtotals only

8.1 Unclear timepoint	5	206	Odds Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]

Comparison 5. FXIII: prophylactic trials: intervention vs inactive control

Ir a la tabla de la revisión

Comparison 6. FXIII: therapeutic trials: intervention vs inactive control

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Allergic Adverse Events Show forest plot	1		Odds Ratio (M‐H, Fixed, 95% CI)	Totals not selected

1.1 Timepoint not stated	1		Odds Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]

Comparison 6. FXIII: therapeutic trials: intervention vs inactive control

Ir a la tabla de la revisión

Comparison 7. PCC: prophylactic trials: intervention vs inactive control

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 All‐cause mortality Show forest plot	1		Risk Ratio (M‐H, Fixed, 95% CI)	Totals not selected

1.1 Time point not stated	1		Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
2 Allergic Adverse Events Show forest plot	1		Odds Ratio (M‐H, Fixed, 95% CI)	Totals not selected

2.1 Timepoint not stated	1		Odds Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]

Comparison 7. PCC: prophylactic trials: intervention vs inactive control

Ir a la tabla de la revisión

	Idioma de la Revisión Cochrane Escoja su idioma de preferencia para las revisiones Cochrane y otros contenidos. Las secciones sin una traducción aparecerán en inglés.

	Idioma de la web Escoja su idioma de preferencia para la web de la Biblioteca Cochrane.

Idioma de la Revisión Cochrane

Idioma de la web

Resumen

Antecedentes

Objetivos

Métodos de búsqueda

Criterios de selección

Obtención y análisis de los datos

Resultados principales

Conclusiones de los autores

PICO

PICO

Population

Intervention

Comparison

Outcome

Resumen en términos sencillos

Efectividad de los fármacos para favorecer la coagulación en la prevención y el tratamiento de la hemorragia en pacientes sin hemofilia

Resumen visual

Authors' conclusions

Implications for practice

Implications for research

Summary of findings

Background

Description of the condition

Description of the intervention

How the intervention might work

Why it is important to do this review

Objectives

Methods

Criteria for considering studies for this review

Types of studies

Types of participants

Types of interventions

Types of outcome measures

Primary outcomes

Secondary outcomes

Search methods for identification of studies

Electronic searches

Searching other resources

Data collection and analysis

Selection of studies

Data extraction and management

Assessment of risk of bias in included studies

Measures of treatment effect

Unit of analysis issues

Dealing with missing data

Method of analysis

Assessment of heterogeneity

Assessment of reporting biases

Data synthesis

Subgroup analysis and investigation of heterogeneity

Sensitivity analysis

'Summary of findings' tables

Results

Description of studies

Results of the search

Included studies

Setting

Participants

Interventions

Prophylactic trials

Therapeutic trials

Co‐interventions

Primary outcomes of included trials

Funding sources

Excluded studies

Risk of bias in included studies

Allocation

Random sequence generation

Prophylactic trials

Therapeutic trials

Concealment of treatment allocation

Prophylactic trials

Therapeutic trials

Blinding

Blinding of participants and personnel (performance bias)

Prophylactic intervention

Therapeutic trials

Blinding of outcome assessors (detection bias)