Methods to decrease blood loss and transfusion requirements for liver transplantation

  • Review
  • Intervention

Authors


Abstract

Background

Excessive blood loss and increased blood transfusion requirements may have significant impact on the short-term and long-term outcomes after liver transplantation.

Objectives

To compare the potential benefits and harms of different methods of decreasing blood loss and blood transfusion requirements during liver transplantation.

Search methods

We searched The Cochrane Central Register of Controlled Trials in The Cochrane Library, MEDLINE, EMBASE, Science Citation Index Expanded, and metaRegister of Controlled Trials until September 2011.

Selection criteria

We included all randomised clinical trials that were performed to compare various methods of decreasing blood loss and blood transfusion requirements during liver transplantation.

Data collection and analysis

Two authors independently identified the trials and extracted the data. We analysed the data with both the fixed-effect and the random-effects model using RevMan Analysis. For each outcome we calculated the risk ratio (RR), mean difference (MD), or standardised mean difference (SMD) with 95% confidence intervals (CI) based on available data analysis. We also conducted network meta-analysis.

Main results

We included 33 trials involving 1913 patients. The sample size in the trials varied from 8 to 209 participants. The interventions included pharmacological interventions (aprotinin, tranexamic acid, epsilon amino caproic acid, antithrombin 3, recombinant factor (rFvIIa), oestrogen, prostaglandin, epinephrine), blood substitutes (blood components rather than whole blood, hydroxy-ethyl starch, thromboelastography), and cardiovascular interventions (low central venous pressure). All the trials were of high risk of bias. Primary outcomes were reported in at least two trials for the following comparisons: aprotinin versus control, tranexamic acid versus control, recombinant factor VIIa (rFVIIa) versus control, and tranexamic acid versus aprotinin. There were no significant differences in the 60-day mortality (3 trials; 6/161 (3.7%) in the aprotinin group versus 8/119 (6.7%) in the control group; RR 0.52; 95% CI 0.18 to 1.45), primary graft non-function (2 trials; 0/128 (0.0%) in the aprotinin group versus 4/89 (4.5%) in the control group; RR 0.15; 95% CI 0.02 to 1.25), retransplantation (3 trials; 2/256 (0.8%) in the aprotinin group versus 12/178 (6.7%) in the control group; RR 0.21; 95% CI 0.02 to 1.79), or thromboembolic episodes (3 trials; 4/161 (2.5%) in the aprotinin group versus 5/119 (4.2%) in the control group; RR 0.59; 95% CI 0.19 to 1.84) between the aprotinin and control groups. There were no significant differences in the 60-day mortality (3 trials; 4/83 (4.8%) in the tranexamic acid group versus 5/56 (8.9%) in the control group; RR 0.55; 95% CI 0.17 to 1.76), retransplantation (2 trials; 3/41 (7.3%) in the tranexamic acid group versus 3/36 (8.3%) in the control group; RR 0.79; 95% CI 0.18 to 3.48), or thromboembolic episodes (5 trials; 5/103 (4.9%) in the tranexamic acid group versus 1/76 (1.3%) in the control group; RR 2.20; 95% CI 0.38 to 12.64) between the tranexamic acid and control groups. There were no significant differences in the 60-day mortality (3 trials; 8/195 (4.1%) in the recombinant factor VIIa (rFVIIa) group versus 2/91 (2.2%) in the control group; RR 1.51; 95% CI 0.33 to 6.95), thromboembolic episodes (2 trials; 24/185 (13.0%) in the rFVIIa group versus 8/81 (9.9%) in the control group; RR 1.38; 95% CI 0.65 to 2.91), or serious adverse events (2 trials; 90/185 (48.6%) in the rFVIIa group versus 30/81 (37.0%) in the control group; RR 1.30; 95% CI 0.94 to 1.78) between the rFVIIa and control groups. There were no significant differences in the 60-day mortality (2 trials; 6/91 (6.6%) in the tranexamic acid group versus 1/87 (1.1%) in the aprotinin group; RR 4.12; 95% CI 0.71 to 23.76) or thromboembolic episodes (2 trials; 4/91 (4.4%) in the tranexamic acid group versus 2/87 (2.3%) in the aprotinin group; RR 1.97; 95% CI 0.37 to 10.37) between the tranexamic acid and aprotinin groups. The remaining outcomes in the above comparisons and the remaining comparisons included only only trial under the primary outcome or the outcome was not reported at all in the trials. There were no significant differences in the mortality, primary graft non-function, graft failure, retransplantation, thromboembolic episodes, or serious adverse events in any of these comparisons. However, the confidence intervals were wide, and it is not possible to reach any conclusion on the safety of the interventions. None of the trials reported the quality of life in patients.

Secondary outcomes were reported in at least two trials for the following comparisons - aprotinin versus control, tranexamic acid versus control, rFVIIa versus control, thromboelastography versus control, and tranexamic acid versus aprotinin. There was significantly lower allogeneic blood transfusion requirements in the aprotinin group than the control group (8 trials; 185 patients in aprotinin group and 190 patients in control group; SMD -0.61; 95% CI -0.82 to -0.40). There were no significant differences in the allogeneic blood transfusion requirements between the tranexamic acid and control groups (4 trials; 93 patients in tranexamic acid group and 66 patients in control group; SMD -0.27; 95% CI -0.59 to 0.06); rFVIIa and control groups (2 trials; 141 patients in rFVIIa group and 80 patients in control group; SMD -0.05; 95% CI -0.32 to 0.23); thromboelastography and control groups (2 trials; 31 patients in thromboelastography group and 31 patients in control group; SMD -0.73; 95% CI -1.69 to 0.24); or between the tranexamic acid and aprotinin groups (3 trials; 101 patients in tranexamic acid group and 97 patients in aprotinin group; SMD -0.09; 95% CI -0.36 to 0.19). The remaining outcomes in the above comparisons and the remaining comparisons included only only trial under the primary outcome or the outcome was not reported at all in the trials. There were no significant differences in the blood loss, transfusion requirements, hospital stay, or intensive care unit stay in most of the comparisons.

Authors' conclusions

Aprotinin, recombinant factor VIIa, and thromboelastography groups may potentially reduce blood loss and transfusion requirements. However, risks of systematic errors (bias) and risks of random errors (play of chance) hamper the confidence in this conclusion. We need further well-designed randomised trials with low risk of systematic error and low risk of random errors before these interventions can be supported or refuted.

Résumé scientifique

Methods to decrease blood loss and transfusion requirements for liver transplantation

Contexte

Les pertes excessives de sang et les besoins accrus de transfusion sanguine peuvent avoir un impact significatif sur les résultats à court terme et à long terme d'une transplantation hépatique.

Objectifs

Comparer les avantages et les inconvénients de différentes méthodes destinées à réduire les pertes sanguines et les besoins de transfusionnels lors d'une transplantation hépatique.

Stratégie de recherche documentaire

Nous avons effectué une recherche dans le registre Cochrane des essais contrôlés de The Cochrane Library, (CENTRAL), dans MEDLINE, EMBASE, Science Citation Index Expanded et dans le méta-registre des essais contrôlés, jusqu'à septembre 2011..

Critères de sélection

Nous avons inclus tous les essais cliniques randomisés qui ont été effectués afin de comparer différentes méthodes pour diminuer les pertes sanguines et les besoins de transfusion sanguine lors d'une transplantation hépatique.

Recueil et analyse des données

Deux auteurs ont identifié les essais et extrait les données de manière indépendante. Nous avons analysé les données avec le modèle à effets fixes et le modèle à effets aléatoires en utilisant le logiciel RevMan. Pour chaque résultat, nous avons calculé le risque relatif (RR), la différence moyenne (DM) ou la différence moyenne standardisée (DMS) avec des intervalles de confiance (IC) à 95 % sur la base de l'analyse des données disponibles. Nous avons également mené une méta-analyse de réseau.

Résultats principaux

Nous avons inclus 33 essais (1913 patients). L’effectif des essais variait de 8 à 209 participants. Les interventions incluaient des interventions pharmacologiques (aprotinine, acide tranexamique, acide epsilon amino-caproique, antithrombine 3, facteur recombinant (rFVIIa), oestrogène, prostaglandine, adrénaline), des substituts du sang (composants sanguins plutôt que sang total, hydroxy-éthyl-amidon, thromboélastographie) et des interventions cardio-vasculaires (basse pression veineuse centrale). Tous les essais étaient à risque élevé de biais. Le critères principaux de jugement étaient rapportés dans au moins deux essais, pour les comparaisons suivantes : aprotinine versus contrôle, acide tranexamique versus contrôle, facteur VIIa recombinant (rFVIIa) versus contrôle et acide tranexamique versus aprotinine. Il n'y avait aucune différence significative entre les groupes à aprotinine et de contrôle en termes de la mortalité à 60 jours (3 essais ; 6/161 (3,7 %) dans le groupe aprotinine contre 8/119 (6,7 %) dans le groupe contrôle ; RR = 0,52 ; IC à 95 % 0,18 à 1,45), de non-fonctionn primaire du greffon (2 essais ; 0/128 (0,0 %) dans le groupe aprotinine contre 4/89 (4,5 %) dans le groupe contrôle ; RR 0,15 ; IC à 95 % 0,02 à 1,25), de retransplantation (3 essais ; 2/256 (0,8 %) dans le groupe aprotinine contre 12/178 (6,7 %) dans le groupe de contrôle ; RR 0,21 ; IC à 95 % 0,02 à 1,79) ou d’épisodes thrombo-emboliques (3 essais ; 4/161 (2,5 %) dans le groupe aprotinine contre 5/119 (4,2 %) dans le groupe de contrôle ; RR = 0,59 ; IC à 95 % 0,19 à 1,84). Il n'y avait aucune différence significative entre les groupes acide tranexamique et contrôle en termes de mortalité à 60 jours (3 essais ; 4/83 (4,8 %) dans le groupe acide tranexamique contre 5/56 (8,9 %) dans le groupe contrôle ; RR 0,55 ; IC à 95 % 0,17 à 1,76), de retransplantation (2 essais ; 3/41 (7,3 %) dans le groupe acide tranexamique contre 3/36 (8,3 %) dans le groupe contrôle ; RR 0,79 ; IC à 95 % 0,18 à 3,48) ou d’épisodes thrombo-emboliques (5 essais ; 5/103 (4,9 %) dans le groupe acide tranexamique contre 1/76 (1,3 %) dans le groupe contrôle ; RR 2,20 ; IC à 95 % 0,38 à 12,64). Il n'y avait aucune différence significative entre les groupes à rFVIIa et contrôle en termes de mortalité à 60 jours (3 essais ; 8/195 (4,1 %) dans le groupe à facteur VIIa recombinant (rFVIIa) contre 2/91 (2,2 %) dans le groupe contrôle ; RR 1,51 ; IC à 95 % 0,33 à 6,95), d’épisodes thrombo-emboliques (2 essais; 24/185 (13,0 %) dans le groupe rFVIIa contre 8/81 (9,9 %) dans le groupe contrôle ; RR 1,38 ; IC à 95 % 0,65 à 2,91) ou d‘événements indésirables graves (2 essais; 90/185 (48,6 %) dans le groupe rFVIIa contre 30/81 (37,0 %) dans le groupe de contrôle ; RR 1,30 ; IC à 95 % 0,94 à 1,78). Il n'y avait aucune différence significative entre les groupes acide tranexamique et aprotinine en termes de mortalité à 60 jours (2 essais ; 6/91 (6,6 % ) dans le groupe acide tranexamique contre 1/87 (1,1 %) dans le groupe aprotinine ; RR 4,12 ; IC à 95 % 0,71 à 23,76) ou d’épisodes thrombo-emboliques (2 essais ; 4/91 (4,4 %) dans le groupe acide tranexamique contre 2/87 (2,3 % dans le groupe aprotinine ; RR 1,97 ; IC à 95 % 0,37 à 10,37). Pour les critères d’evaluaiton résultants des comparaisons ci-dessus, ils n’etaient disponibles que dans un seul, ou aucun essai. Pour aucune de ces comparaisons il n'y avait pas de différence significative en termes de mortalité, de non-fonction primaire du greffon, d'échec de la greffe, de retransplantation, d’épisodes thrombo-emboliques ou d’événements indésirables graves. Cependant, les intervalles de confiance étaient larges et il n'est pas possible de tirer quelque conclusion que ce soit sur l'innocuité des interventions. Aucun des essais n'avait rapporté des données de qualité de vie des patients..

Les critères de jugement secondaires étaient rapportés pour les comparaisons suivantes : - aprotinine versus contrôle, acide tranexamique versus contrôle, rFVIIa versus contrôle, thromboélastographie versus contrôle et acide tranexamique versus aprotinine. Il y avait des besoins significativement plus faibles en transfusion sanguine allogénique dans le groupe à aprotinine que dans le groupe contrôle (8 essais ; 185 patients dans le groupe aprotinine et 190 patients dans le groupe de contrôle ; SMD -0,61 ; IC à 95 % -0,82 à -0,40). Il n'y avait pas de différences significatives en besoins en transfusion sanguine allogénique entre les groupes acide tranexamique et contrôles (4 essais ; 93 patients dans le groupe acide tranexamique et 66 patients dans le groupe contrôle ; SMD -0,27 ; IC à 95 % -0,59 à 0,06), entre les groupes rFVIIa et contrôle (2 essais ; 141 patients dans le groupe rFVIIa et 80 patients dans le groupe contrôle ; SMD -0,05 ; IC à 95 % -0,32 à 0,23), entre les groupes thromboélastographie et contrôle (2 essais ; 31 patients dans le groupe thromboélastographie et 31 patients dans le groupe de contrôle ; SMD -0,73 ; IC à 95 % -1,69 à 0,24) ou entre les groupes acide tranexamique et aprotinine (3 essais ; 101 patients dans le groupe acide tranexamique et 97 patients dans le groupe aprotinine ; SMD -0,09 ; IC à 95 % -0,36 à 0,19). Pour les autres résultats resultants des comparaisons ci-dessus, il n'en était fait état que dans un seul essai. Pour ce qui concerne les autres comparaisons, il n'était fait état des principaux résultats que dans un seul essai au maximum. Pour la plupart des comparaisons il n'y avait aucune différence significative en termes de pertes sanguines, de besoins transfusionnels, de durée du séjour à l'hôpital ou de séjour en unité de soins intensifs.

Conclusions des auteurs

L'aprotinine, le facteur VIIa recombinant et la thromboélastographie sont susceptibles de réduire les pertes sanguines et les besoins transfusionnels. Cependant, les risques d'erreurs systématiques (biais) et les risques d'erreurs aléatoires (effet de hasard) entravent la confiance dans cette conclusion. Nous avons besoin d'essais randomisés supplémentaires, bien conçus, avec un faible risque d'erreur systématique et un faible risque d'erreurs aléatoires, avant que ces interventions puissent être validées ou récusées.

Resumen

Métodos para reducir la pérdida de sangre y la necesidad de transfusión en el trasplante hepático

Antecedentes

La pérdida sanguínea excesiva y el aumento de la necesidad de transfusión de sangre pueden tener una repercusión significativa sobre los resultados a corto plazo y a largo plazo después del trasplante hepático.

Objetivos

Comparar los efectos beneficiosos y perjudiciales potenciales de los diferentes métodos para reducir la pérdida de sangre y la necesidad de transfusión durante el trasplante hepático.

Métodos de búsqueda

Se hicieron búsquedas en el Registro Cochrane Central de Ensayos Controlados (Cochrane Central Register of Controlled Trials) en The Cochrane Library, MEDLINE, EMBASE, Science Citation Index Expanded y en el metaRegister of Controlled Trials hasta septiembre 2011.

Criterios de selección

Se incluyeron todos los ensayos clínicos aleatorios realizados para comparar varios métodos para reducir la pérdida de sangre y la necesidad de transfusión de sangre durante el trasplante hepático.

Obtención y análisis de los datos

Dos autores, de forma independiente, identificaron los ensayos y extrajeron los datos. Se analizaron los datos con el modelo de efectos fijos y el modelo de efectos aleatorios con el uso del análisis en RevMan. Para cada resultado, se calculó el cociente de riesgos (CR), la diferencia de medias (DM) o la diferencia de medias estandarizada (DME) con intervalos de confianza (IC) del 95%, según el análisis de datos disponibles. También se realizó un metanálisis de redes.

Resultados principales

Se incluyeron 33 ensayos con 1913 pacientes. El tamaño de muestra en los ensayos varió de 8 a 209 participantes. Las intervenciones incluyeron intervenciones farmacológicas (aprotinina, ácido tranexámico, ácido épsilon aminocaproico, antitrombina 3, factor recombinante (rFVIIa), estrógeno, prostaglandina, epinefrina), sustitutos sanguíneos (componentes sanguíneos en lugar de sangre total, almidón hidroxietilo, tromboelastografía) e intervenciones cardiovasculares (presión venosa central baja). Todos los ensayos presentaron un alto riesgo de sesgo. Los resultados primarios se informaron en al menos dos ensayos para las siguientes comparaciones: aprotinina versus control, ácido tranexámico versus control, factor VIIa recombinante (rFVIIa) versus control, y ácido tranexámico versus aprotinina. No hubo diferencias significativas en la mortalidad a los 60 días (tres ensayos; 6/161 [3,7%] en el grupo de aprotinina versus 8/119 [6,7%] en el grupo control; CR 0,52; IC del 95%: 0,18 a 1,45), el injerto primario no funcionante (dos ensayos; 0/128 [0,0%] en el grupo de aprotinina versus 4/89 [4,5%] en el grupo control; CR 0,15; IC del 95%: 0,02 a 1,25), el nuevo trasplante (tres ensayos; 2/256 [0,8%] en el grupo de aprotinina versus 12/178 [6,7%] en el grupo control; CR 0,21; IC del 95%: 0,02 a 1,79), ni en los episodios tromboembólicos (tres ensayos; 4/161 [2,5%] en el grupo de aprotinina versus 5/119 [4,2%] en el grupo control; CR 0,59; IC del 95%: 0,19 a 1,84) entre los grupos aprotinina y control. No hubo diferencias significativas en la mortalidad a los 60 días (tres ensayos; 4/83 [4,8%] en el grupo de ácido tranexámico versus 5/56 [8,9%] en el grupo control; CR 0,55; IC del 95%: 0,17 a 1,76), el nuevo trasplante (dos ensayos; 3/41 [7,3%] en el grupo de ácido tranexámico versus 3/36 [8,3%] en el grupo control; CR 0,79; IC del 95%: 0,18 a 3,48), ni en los episodios tromboembólicos (cinco ensayos; 5/103 [4,9%] en el grupo de ácido tranexámico versus 1/76 [1,3%] en el grupo control; CR 2,20; IC del 95%: 0,38 a 12,64) entre los grupos ácido tranexámico y control. No hubo diferencias significativas en la mortalidad a los 60 días (tres ensayos; 8/195 [4,1%] en el grupo de factor VIIa recombinante [rFVIIa] versus 2/91 [2,2%] en el grupo control; CR 1,51; IC del 95%: 0,33 a 6,95), los episodios tromboembólicos (dos ensayos; 24/185 [13,0%] en el grupo rFVIIa versus 8/81 [9,9%] en el grupo control; CR 1,38; IC del 95%: 0,65 a 2,91), ni en los eventos adversos graves (dos ensayos; 90/185 [48,6%] en el grupo rFVIIa versus 30/81 [37,0%] en el grupo control; CR 1,30; IC del 95%: 0,94 a 1,78) entre los grupos rFVIIa y control. No hubo diferencias significativas en la mortalidad a los 60 días (dos ensayos; 6/91 [6,6%] en el grupo ácido tranexámico versus 1/87 [1,1%] en el grupo aprotinina; CR 4,12; IC del 95%: 0,71 a 23,76) ni en los episodios tromboembólicos (dos ensayos; 4/91 [4,4%] en el grupo ácido tranexámico versus 2/87 [2,3%] en el grupo aprotinina; CR 1,97; IC del 95%: 0,37 a 10,37) entre los grupos ácido tranexámico y aprotinina. Los resultados restantes en las comparaciones anteriores y las comparaciones restantes solo incluyeron un ensayo bajo el resultado primario o el resultado no se informó en absoluto en los ensayos. No hubo diferencias significativas en la mortalidad, el injerto primario no funcionante, el fracaso del injerto, el nuevo trasplante, los episodios tromboembólicos ni los eventos adversos graves en cualquiera de estas comparaciones. Sin embargo, los intervalos de confianza fueron amplios y no es posible establecer ninguna conclusión sobre la seguridad de las intervenciones. Ninguno de los ensayos informó la calidad de vida en los pacientes.

Los resultados secundarios se informaron en al menos dos ensayos para las siguientes comparaciones: aprotinina versus control, ácido tranexámico versus control, rFVIIa versus control, tromboelastografía versus control y ácido tranexámico versus aprotinina. Hubo una necesidad de transfusión de sangre alogénica significativamente menor en el grupo de aprotinina que en el grupo control (ocho ensayos; 185 pacientes del grupo de aprotinina y 190 pacientes del grupo control; DME -0,61; IC del 95%: -0,82 a -0,40). No hubo diferencias significativas en la necesidad de transfusión de sangre alogénica entre los grupos ácido tranexámico y control (cuatro ensayos; 93 pacientes del grupo de ácido tranexámico y 66 pacientes del grupo control; DME -0,27; IC del 95%: -0,59 a 0,06); los grupos rFVIIa y control (dos ensayos; 141 pacientes del grupo de rFVIIa y 80 pacientes del grupo control; DME -0,05; IC del 95%: -0,32 a 0,23); los grupos tromboelastografía y control (dos ensayos; 31 pacientes del grupo de tromboelastografía y 31 pacientes del grupo control; DME -0,73; IC del 95%: -1,69 a 0,24); ni entre los grupos ácido tranexámico y aprotinina (tres ensayos; 101 pacientes del grupo de ácido tranexámico y 97 del grupo de aprotinina; DME -0,09; IC del 95%: -0,36 a 0,19). Los resultados restantes en las comparaciones anteriores y las comparaciones restantes solo incluyeron un ensayo bajo el resultado primario o el resultado no se informó en absoluto en los ensayos. No hubo diferencias significativas en la pérdida sanguínea, la necesidad de transfusión, la estancia hospitalaria ni la estancia en la unidad de cuidados intensivos en la mayoría de las comparaciones.

Conclusiones de los autores

Los grupos aprotinina, factor VIIa recombinante y tromboelastografía pueden reducir potencialmente la pérdida sanguínea y la necesidad de transfusión. Sin embargo, los riesgos de errores sistemáticos (sesgo) y los riesgos de errores aleatorios (influencia del azar) disminuyen la confianza en esta conclusión. Se necesitan ensayos aleatorios adicionales bien diseñados con bajo riesgo de errores sistemáticos y aleatorios antes de poder apoyar o refutar estas intervenciones.

Plain language summary

Methods to decrease blood loss and transfusion requirements for liver transplantation

The liver is the powerhouse of the body. It acts as a store of energy and a centre of metabolic activity. Liver transplantation is the main treatment for severe liver disease resulting in destruction of the liver (which can happen suddenly or over a period of time) due to various causes including alcoholism, viral infections, and autoimmune diseases. Liver transplantation is a major surgical procedure and is associated with significant loss of blood. Various methods have been used to decrease blood loss and transfusion requirements in patients undergoing liver transplantation, with a view to improve the results of liver transplantation. We performed a detailed review of the medical literature (available until September 2011) to determine the benefits and harms of different methods of decreasing blood loss and transfusion requirements in patients undergoing liver transplantation. We sought evidence from randomised clinical trials only, as when conducted properly such studies provide the best evidence. Two authors independently identified the trials and obtained the information from the trials.

We included 33 trials involving 1913 patients. The number of patients included in the trials varied from 8 to 209. The comparisons included various drugs that affect the blood clotting (congealing) such as aprotinin, tranexamic acid; blood substitutes (blood components rather than whole blood); use of thromboelastography (a bedside measure of blood clot formation); and lowering the pressure in the veins with an aim to decrease the blood loss from veins. We found no significant difference in the risk of death or graft loss, or in the major complication rates between the compared groups in any of the comparisons. Quality of life was not reported in any of the trials. There does not appear to be any consistency in the results between blood loss and blood transfusion requirements. Aprotinin, tranexamic acid, recombinant factor VIIa, low central venous pressure, and thromboelastography may lower blood loss and transfusion requirements. However, these findings are based on few trials with a high risk of bias (systematic overestimation of benefits) and high risk of play of chance (random error due to small number of patients). There were no differences in the hospital stay or intensive care unit stay in any of the comparisons. Nor was there any significant difference in the intensive therapy unit stay, or hospital stay between the compared groups. Again, most of the trials were of high risk of systematic errors (a potential to arrive at wrong conclusions because of the way the trial was conducted) and random errors (a potential to arrive at wrong conclusions because of play of chance).

Aprotinin is a drug which has been withdrawn from market since there was a suspicion that it increased death after major heart operations. The results from this review do not reveal any increased mortality with aprotinin in the liver transplantation setting although one has to interpret this information with caution because of the few patients included in the trial. We are unable to advocate or refute any method of decreasing blood loss and transfusion requirements in patients undergoing liver transplantation. Further well designed trials with low risk of systematic error and low risk of random errors are necessary.

Résumé simplifié

Méthodes pour diminuer les pertes sanguines et les besoins transfusionnels dans la transplantation du foie

Le foie est la centrale énergétique du corps. Il agit comme une réserve d'énergie et comme centre d'activité métabolique. La transplantation du foie est le principal traitement pour les maladies grave du foie entraînant sa destruction (qui peut se produire soudainement ou s'étaler dans le temps) et qui est due à des causes diverses comme l'alcoolisme, les infections virales et les maladies auto-immunes. La transplantation du foie est une procédure chirurgicale majeure et elle est associée à une perte importante de sang. Diverses méthodes ont été utilisées pour diminuer les pertes sanguines et les besoins transfusionnels chez les patients subissant une transplantation du foie, en vue d'améliorer les résultats de la transplantation. Nous avons effectué une revue détaillée de la littérature médicale (disponible jusqu'en septembre 2011) pour déterminer les bénéfices et les inconvénients de différentes méthodes de réduction des pertes sanguines et des besoins transfusionnels chez les patients subissant une transplantation du foie. Nous n'avons cherché de données provenant des essais cliniques randomisés, parce que lorsqu'elles sont menées comme il faut ces études sont celles qui fournissent les données les plus fiables. Deux auteurs ont indépendamment identifié les essais et obtenu les informations des essais.

Nous avons inclus 33 essais (1 913 patients). Le nombre de patients inclus dans les essais variait entre 8 et 209. Les comparaisons comprenaient : divers médicaments affectant la coagulation sanguine (aprotinine, acide tranexamique..), des substituts du sang (des composants sanguins plutôt que le sang total), l'utilisation de la thromboélastographie (une mesure au chevet de la formation de caillots sanguins) et l'abaissement de la pression dans les veines dans le but de diminuer les fuites de sang. Nous n'avons trouvé aucune différence significative concernant le risque de décès ou de perte du greffon, ou les taux de complications majeures entre les groupes comparés dans une quelconque comparaison. Dans aucun des essais la qualité de vie a été etudiée. Il ne semble pas y avoir de cohérence entre les résultats de pertes de sang et de besoins en transfusion. L'aprotinine, l'acide tranexamique, le facteur VIIa recombinant, la faible pression veineuse centrale et la thromboélastographie sont susceptibles de réduire les pertes de sang et les besoins en transfusion. Cependant, ces résultats sont basés sur quelques essais à risque élevé de biais (surestimation systématique des bénéfices) et à risque élevé d'effet de hasard (erreur aléatoire due au petit nombre de patients). Il n'y avait pas de différences quant à la durée d'hospitalisation ou de risque de séjour en soins intensifs dans aucune des comparaisons. Il n'y avait pas non plus de différence significative entre les groupes comparés pour la durée du séjour en unité de soins intensifs ou à l'hôpital. La plupart des essais présentaient un risque élevé d'erreurs systématiques (un potentiel de parvenir à de mauvaises conclusions en raison de la manière dont l'essai était mené) et d'erreurs aléatoires (un potentiel de parvenir à de mauvaises conclusions par l'effet du hasard).

L'aprotinine est un médicament qui a été retiré du marché depuis que l'on soupçonne qu'il augmente le taux de décès après les opérations cardiaques majeures. Les résultats de cette revue ne révèlent aucune augmentation de la mortalité avec l'aprotinine dans un contexte de transplantation du foie, bien qu'il faille interpréter cette information avec prudence en raison du petit nombre de patients inclus dans l'essai. Nous ne sommes en mesure de valider ou de récuser aucune méthode de réduction des pertes sanguines et des besoins transfusionnels chez les patients subissant une transplantation du foie. Il est nécessaire de mener d'autres essais bien conçus à faible risque d'erreurs systématiques et à faible risque d'erreurs aléatoires.

Notes de traduction

Traduit par: French Cochrane Centre 1st January, 2012
Traduction financée par: Ministère du Travail, de l'Emploi et de la Santé Français

Resumen en términos sencillos

Métodos para reducir la pérdida de sangre y la necesidad de transfusión en el trasplante hepático

El hígado es la fuente de energía del cuerpo. Actúa como una reserva de energía y un centro de la actividad metabólica. El trasplante hepático es el tratamiento principal para las enfermedades hepáticas graves que dan lugar a la destrucción del hígado (que puede suceder repentinamente o durante cierto período de tiempo) debido a diversas causas que incluyen el alcoholismo, las infecciones virales y las enfermedades autoinmunes. El trasplante hepático es un procedimiento quirúrgico mayor y se asocia con pérdida significativa de sangre. Se han utilizado diversos métodos para disminuir la pérdida sanguínea y la necesidad de transfusión en los pacientes a los que se les realiza trasplante hepático con el objetivo de mejorar los resultados del trasplante hepático. Se realizó una revisión detallada de la bibliografía médica (disponible hasta setiembre de 2011) para determinar los efectos beneficiosos y perjudiciales de diferentes métodos para reducir la pérdida sanguínea y la necesidad de transfusión en los pacientes a los que se les realiza trasplante hepático. Se buscaron pruebas de ensayos clínicos aleatorios solamente, ya que cuando se realizan de forma apropiada dichos estudios aportan las mejores pruebas. Dos autores, de forma independiente, identificaron los ensayos y obtuvieron la información de los mismos.

Se incluyeron 33 ensayos con 1913 pacientes. El número de pacientes incluidos en los ensayos varió desde ocho a 209. Las comparaciones incluyeron diversos fármacos que afectan la coagulación de la sangre (espesamiento) como la aprotinina, el ácido tranexámico; los sustitutos sanguíneos (componentes de la sangre en lugar de sangre total); el uso de la tromboelastografía (una medida de cabecera de formación de coágulos sanguíneos); y la disminución de la presión en las venas con el objetivo de reducir la pérdida sanguínea de las venas. No se encontraron diferencias significativas en el riesgo de muerte ni en la pérdida del injerto, ni en las tasas de complicaciones importantes entre los grupos comparados en cualquiera de las comparaciones. En ninguno de los ensayos se informó la calidad de vida. Al parecer, no hay consistencia en los resultados entre la pérdida sanguínea y la necesidad de transfusión de sangre. La aprotinina, el ácido tranexámico, el factor VIIa recombinante, la presión venosa central baja y la tromboelastografía pueden reducir la pérdida sanguínea y la necesidad de transfusión. Sin embargo, estos hallazgos se basan en pocos ensayos con un alto riesgo de sesgo (sobrestimación sistemática de los efectos beneficiosos) y alto riesgo de influencia del azar (error aleatorio debido al escaso número de pacientes). No hubo diferencias en la estancia hospitalaria ni en la estancia en la unidad de cuidados intensivos en cualquiera de las comparaciones. Tampoco hubo diferencias significativas en la estancia en la unidad de terapia intensiva ni en la estancia hospitalaria entre los grupos comparados. Nuevamente, la mayoría de los ensayos tuvo un alto riesgo de errores sistemáticos (la posibilidad de establecer conclusiones equivocadas debido a la manera en la que se realizó el ensayo) y errores aleatorios (la posibilidad de establecer conclusiones equivocadas debido a la influencia del azar).

La aprotinina es un fármaco que se retiró del mercado desde que surgió la sospecha de que aumentó la muerte después de operaciones cardíacas mayores. Los resultados de esta revisión no muestran un aumento de la mortalidad con la aprotinina en el ámbito del trasplante hepático, aunque es necesario interpretar esta información con precaución debido a los pocos pacientes incluidos en los ensayos. No es posible recomendar ni refutar cualquier método de reducción de la pérdida sanguínea y la necesidad de transfusión en los pacientes a los que se les realiza trasplante hepático. Se necesitan ensayos adicionales bien diseñados con bajo riesgo de errores sistemáticos y bajo riesgo de errores aleatorios.

Notas de traducción

Traducido por: Centro Cochrane Iberoamericano
Traducción patrocinada por: No especificada

Summary of findings(Explanation)

Summary of findings for the main comparison. Intervention versus control for liver transplantation (mortality, primary graft non-function, and retransplantation)
  1. 1 All trials were at high risk of bias,
    2 The confidence intervals overlap 0.75 and 1.25.
    3 Funnel plots could not be performed for any of the outcomes.
    4 Although the confidence intervals do not overlap 0.75 and 1.25, the confidence intervals were wide.

Intervention versus control for liver transplantation
Patient or population: Patients with liver transplantation.
Settings: Transplantation centre.
Intervention: Intervention versus control.
OutcomesIllustrative comparative risks* (95% CI)Relative effect
(95% CI)
No of Participants
(studies)
Quality of the evidence
(GRADE)
Comments
Assumed riskCorresponding risk
ControlIntervention versus control
60-day mortality - Aprotinin versus controlStudy populationRR 0.52
(0.18 to 1.45)
280
(3 studies)
⊕⊝⊝⊝
very low 1,2,3
 
67 per 100035 per 1000
(12 to 97)
Moderate
73 per 100038 per 1000
(13 to 106)
60-day mortality - Tranexamic acid versus controlStudy populationRR 0.55
(0.17 to 1.76)
139
(3 studies)
⊕⊝⊝⊝
very low 1,2,3
 
89 per 100049 per 1000
(15 to 157)
Moderate
100 per 100055 per 1000
(17 to 176)
60-day mortality - Recombinant factor VIIa (rFVIIa) versus controlStudy populationRR 1.51
(0.33 to 6.95)
286
(3 studies)
⊕⊝⊝⊝
very low 1,2,3
 
22 per 100033 per 1000
(7 to 153)
Moderate
16 per 100024 per 1000
(5 to 111)
60-day mortality - Tranexamic acid versus aprotininStudy populationRR 4.12
(0.71 to 23.76)
178
(2 studies)
⊕⊝⊝⊝
very low 1,2,3
 
11 per 100047 per 1000
(8 to 273)
Moderate
8 per 100033 per 1000
(6 to 190)
Primary graft non-function - Aprotinin versus controlStudy populationRR 0.15
(0.02 to 1.25)
217
(2 studies)
⊕⊝⊝⊝
very low 1,3,4
 
45 per 10007 per 1000
(1 to 56)
Moderate
45 per 10007 per 1000
(1 to 56)
Retransplantation - Aprotinin versus controlStudy populationRR 0.21
(0.02 to 1.79)
217
(2 studies)
⊕⊝⊝⊝
very low 1,3,4
 
67 per 100014 per 1000
(1 to 121)
Moderate
66 per 100014 per 1000
(1 to 118)
Retransplantation - Tranexamic acid versus controlStudy populationRR 0.79
(0.18 to 3.48)
77
(2 studies)
⊕⊝⊝⊝
very low 1,2,3
 
83 per 100066 per 1000
(15 to 290)
Moderate
75 per 100059 per 1000
(14 to 261)
*The basis for the assumed risk (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI).
CI: Confidence interval; RR: Risk ratio.
GRADE Working Group grades of evidence
High quality: Further research is very unlikely to change our confidence in the estimate of effect.
Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate.
Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate.
Very low quality: We are very uncertain about the estimate.

Summary of findings 2 Intervention versus control for liver transplantation (thromboembolic episodes and other serious adverse events)

Summary of findings 2. Intervention versus control for liver transplantation (thromboembolic episodes and other serious adverse events)
  1. 1 All trials were at high risk of bias,
    2 The confidence intervals overlap 0.75 and 1.25.
    3 Funnel plots could not be performed for any of the outcomes.
    4 Although the confidence intervals do not overlap 0.75 and 1.25, the confidence intervals were wide.

Intervention versus control for liver transplantation
Patient or population: Patients with liver transplantation.
Settings: Transplantation centre.
Intervention: Intervention versus control.
OutcomesIllustrative comparative risks* (95% CI)Relative effect
(95% CI)
No of Participants
(studies)
Quality of the evidence
(GRADE)
Comments
Assumed riskCorresponding risk
ControlIntervention versus control
Thromboembolic episodes - Aprotinin versus controlStudy populationRR 0.6
(0.18 to 1.96)
280
(3 studies)
⊕⊝⊝⊝
very low 1,2,3
 
42 per 100025 per 1000
(8 to 82)
Moderate
63 per 100038 per 1000
(11 to 123)
Thromboembolic episodes - Tranexamic acid versus controlStudy populationRR 2.2
(0.38 to 12.64)
179
(5 studies)
⊕⊝⊝⊝
very low 1,2,3
 
13 per 100029 per 1000
(5 to 166)
Thromboembolic episodes - Recombinant factor VIIa (rFVIIa) versus controlStudy populationRR 1.38
(0.65 to 2.91)
266
(2 studies)
⊕⊝⊝⊝
very low 1,2,3
 
99 per 1000136 per 1000
(64 to 287)
Moderate
101 per 1000139 per 1000
(66 to 294)
Serious adverse events - Recombinant factor VIIa (rFVIIa) versus controlStudy populationRR 1.3
(0.94 to 1.78)
266
(2 studies)
⊕⊝⊝⊝
very low 1,3,4
 
370 per 1000481 per 1000
(348 to 659)
Moderate
406 per 1000528 per 1000
(382 to 723)
*The basis for the assumed risk (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI).
CI: Confidence interval; RR: Risk ratio.
GRADE Working Group grades of evidence
High quality: Further research is very unlikely to change our confidence in the estimate of effect.
Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate.
Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate.
Very low quality: We are very uncertain about the estimate.

Summary of findings 3 Intervention versus control for liver transplantation (blood loss)

Summary of findings 3. Intervention versus control for liver transplantation (blood loss)
  1. 1 All trials were at high risk of bias,
    2 Sample size was less than 400 patients in both groups put together,
    3 Funnel plots could not be performed for any of the outcomes.

Intervention versus control for liver transplantation
Patient or population: Patients with liver transplantation.
Settings: Transplantation centre.
Intervention: Intervention versus control.
OutcomesIllustrative comparative risks* (95% CI)Relative effect
(95% CI)
No of Participants
(studies)
Quality of the evidence
(GRADE)
Comments
Assumed riskCorresponding risk
ControlIntervention versus control
Blood loss - Aprotinin versus control (litre) The mean blood loss - aprotinin versus control in the intervention groups was
1.36 lower
(3.39 lower to 0.66 higher)
 195
(3 studies)
⊕⊝⊝⊝
very low 1,2,3
 
Blood loss - Tranexamic acid versus control (litre) The mean blood loss - tranexamic acid versus control in the intervention groups was
4.98 lower
(10.18 lower to 0.23 higher)
 65
(2 studies)
⊕⊝⊝⊝
very low 1,2,3
 
Blood loss - Thromboelastography versus control (litre) The mean blood loss - thromboelastography versus control in the intervention groups was
1.13 lower
(1.85 to 0.41 lower)
 62
(2 studies)
⊕⊝⊝⊝
very low 1,2,3
 
Blood loss - Tranexamic acid versus aprotinin The mean blood loss - tranexamic acid versus aprotinin in the intervention groups was
1.01 lower
(2.31 lower to 0.29 higher)
 71
(2 studies)
⊕⊝⊝⊝
very low 1,2,3
 
*The basis for the assumed risk (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI).
CI: Confidence interval.
GRADE Working Group grades of evidence
High quality: Further research is very unlikely to change our confidence in the estimate of effect.
Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate.
Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate.
Very low quality: We are very uncertain about the estimate.

Summary of findings 4 Intervention versus control for liver transplantation (red cell or whole blood allogeneic transfusion)

Summary of findings 4. Intervention versus control for liver transplantation (red cell or whole blood allogeneic transfusion)
  1. 1 All trials were at high risk of bias.
    2 High heterogeneity.
    3 Funnel plots could not be performed for any of the outcomes.
    4 Sample size was less than 400 patients in both groups put together.

Intervention versus control for liver transplantation
Patient or population: Patients with liver transplantation.
Settings: Transplantation centre.
Intervention: Intervention versus control.
OutcomesIllustrative comparative risks* (95% CI)Relative effect
(95% CI)
No of Participants
(studies)
Quality of the evidence
(GRADE)
Comments
Assumed riskCorresponding risk
ControlIntervention versus control
Red cell or whole blood transfusion - Aprotinin versus control The mean red cell or whole blood transfusion - aprotinin versus control in the intervention groups was
0.61 standard deviations lower
(0.82 to 0.40 lower)
 375
(8 studies)
⊕⊝⊝⊝
very low 1,2,3
SMD -0.61 (-0.82 to -0.40)
Red cell or whole blood transfusion - Tranexamic acid versus control The mean red cell or whole blood transfusion - tranexamic acid versus control in the intervention groups was
0.27 standard deviations lower
(0.59 lower to 0.06 higher)
 159
(4 studies)
⊕⊝⊝⊝
very low 1,3,4
SMD -0.27 (-0.59 to 0.06)
Red cell or whole blood transfusion - Recombinant factor VIIa (rFVIIa) versus control The mean red cell or whole blood transfusion - recombinant factor VIIa (rFVIIa) versus control in the intervention groups was
0.05 standard deviations higher
(0.32 lower to 0.23 higher)
 221
(2 studies)
⊕⊝⊝⊝
very low 1,3,4
SMD 0.05 (-0.32 to 0.23)
Red cell or whole blood transfusion - Thromboelastography versus control The mean red cell or whole blood transfusion - thromboelastography versus control in the intervention groups was
0.73 standard deviations lower
(1.69 lower to 0.24 higher)
 62
(2 studies)
⊕⊝⊝⊝
very low 1,2,3,4
SMD -0.73 (-1.69 to 0.24)
Red cell or whole blood transfusion - Tranexamic acid versus aprotinin The mean red cell or whole blood transfusion - tranexamic acid versus aprotinin in the intervention groups was
0.09 standard deviations lower
(0.36 lower to 0.19 higher)
 198
(3 studies)
⊕⊝⊝⊝
very low 1,3,4
SMD -0.09 (-0.36 to 0.19)
*The basis for the assumed risk (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI).
CI: Confidence interval.
GRADE Working Group grades of evidence
High quality: Further research is very unlikely to change our confidence in the estimate of effect.
Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate.
Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate.
Very low quality: We are very uncertain about the estimate.

Background

Description of the condition

Liver transplantation is the replacement of a diseased recipient liver with a donor liver. Annually around 5700 liver transplantations are performed in Europe (ELTR 2011) and around 6000 in the United States of America (OPTN/SRTR 2009). Liver transplantation is performed mainly for end-stage liver failure arising as a result of chronic liver disease (for example, alcoholic cirrhosis or viral disease), for acute liver failure (for example viral or due to drug overdose), or for tumour (Lim 2006). The model for end-stage liver disease score (MELD score) has been suggested as one of the methods of determining the severity of end-stage liver failure (Kamath 2001) and is being used as a tool for allocation of donor livers in some countries (Shiffman 2006). Worldwide, there is a demand for liver transplants in surplus of supply. Liver grafts are mainly harvested from cadavers (Koneru 2005; Cescon 2006), but due to shortage of cadaveric organs, there is an increasing interest in live donors who donate part of their liver to the recipients (Jeon 2010). Currently, around 4% of liver grafts are retrieved from live donors (OPTN/SRTR 2009). Recently, split liver transplantation (using one cadaveric donor liver for two recipients, ie, an adult and a paediatric recipient) has been suggested as a way to decrease the organ shortage for liver transplant (Corno 2006).

During liver transplantation, the diseased liver has to be removed. Because of portal hypertension and resulting variceal blood vessels as a consequence of end-stage liver failure, blood loss during removal of the diseased liver can be high. In addition, because of clotting disorders related to liver failure, the blood loss during liver transplantation can be high. Peri-operative blood loss is one of the causes of mortality due to liver transplantation (Bismuth 1987). In observational studies, differences in blood transfusion requirements are associated with increased short-term morbidity such as infective episodes (Nardo 2005), re-operations (Hendriks 2005), and differences in the long- term patient survival (Ramos 2003; Massicotte 2005; Boyd 2007; Boin 2008) in patients undergoing liver transplantation.

Description of the intervention

Various methods have been attempted with an aim to decrease the blood loss and allogeneic blood transfusion requirements during liver transplantation. These include interventions such as lowering central venous pressure (CVP) (Massicotte 2006); acute normovolemic haemodilution (Jabbour 2005); intra-operative blood salvage (Sankarankutty 2006); thromboelastography (Wang 2010), and the use of pharmacological agents such as antifibrinolytics (aprotinin, tranexamic acid, epsilon amino caproic acid (EACA)) (Dalmau 2000; Dalmau 2004); prostaglandins (Himmelreich 1993); conjugated oestrogens (Frenette 1998); desmopressin (Pivalizza 2003); antithrombin III (Palareti 1991); or recombinant factor VIIa (Lodge 2005).

How the intervention might work

Different interventions work in different ways. Lowering central venous pressure aims to decrease the venous blood loss. Acute normovolemic haemodilution aims to decrease the blood loss by diluting the blood. Dilution of blood reduces blood transfusion requirements by decreasing loss of blood constituents, this means that even if the same volume of blood is lost, the proportion of this blood that requires replacing by transfusion is less. Intra-operative blood salvage aims to recover the blood lost by the patient, process it, and transfuse the patient the same blood thereby decreasing the need for allogeneic blood (donated blood). Thromboelastography provides rapid information about global clotting and is performed in the operation theatre (point-of-care approach). It provides rapid results as opposed to laboratory results, which may not be available immediately (Wang 2010). Various pharmacological agents are aimed at decreasing the blood loss and transfusion requirements by altering the coagulation cascade.

Why it is important to do this review

Previous systematic reviews mainly assessed the role of antifibrinolytics and concluded that antifibrinolytics (mainly aprotinin) were safe and decreased blood loss but these studies have failed to demonstrate a reduction in the overall mortality or morbidity (Molenaar 2007; Liu 2008). There has been no Cochrane review assessing the different methods, specified above, to decrease blood loss for liver transplantation.

Objectives

To compare the benefits and harms of different intervention methods aimed at decreasing blood loss and/or allogeneic blood transfusion requirements during liver transplantation.

Methods

Criteria for considering studies for this review

Types of studies

We included randomised clinical trials irrespective of blinding, language, publication status, or sample size. We planned to exclude quasi-randomised trials (eg, allocation by date of birth, day of the week, etc) with regards to the benefits of the intervention but include such studies with regards to harms.

Types of participants

Patients undergoing orthotopic liver transplantation irrespective of age, living donor or cadaveric donor, and reason for liver transplantation.

Types of interventions

The following comparisons were included.

  • Any intervention aimed at decreasing blood loss and/or allogeneic blood transfusion in recipient operation during liver transplantation versus no intervention or placebo.

  • Any intervention aimed at decreasing blood loss and/or allogeneic blood transfusion in recipient operation during liver transplantation versus the same intervention at a different dose or another intervention.

Interventions in living donor retrieval will not be included as they were assessed in other two Cochrane reviews (Gurusamy 2009a; Gurusamy 2009b).

Types of outcome measures

Primary outcomes
  1. Mortality.

    1. 60-day mortality.

    2. At maximal follow-up.

  2. Graft failure/retransplantation.

    1. Primary graft non-function/failure.

    2. Graft failure at maximal follow-up.

    3. Retransplantation for any cause.

  3. Adverse events defined as any untoward medical occurrence not necessarily having a causal relationship with the treatment, but resulting in a dose reduction or discontinuation of treatment (ICH-GCP 1997). Serious adverse events are defined as any event that would increase mortality; is life-threatening; requires inpatient hospitalisation; results in a persistent or significant disability; or any important medical event, which might have jeopardised the patient or requires intervention to prevent it. Any surgery related morbidity such as re-operation for any cause, sepsis, bile leak, renal failure, wound complications, chest infection are likely to result in prolongation of inpatient hospitalisation and were classified as serious adverse events. We obtained information for thromboembolic events such as hepatic artery thrombosis and portal vein thrombosis separately, since many of the interventions affect the coagulation cascade, which may result in increased thromboembolic episodes.

  4. Quality of life.

Secondary outcomes
  1. Blood loss.

    1. Blood loss (however measured by authors).

    2. Allogeneic transfusion requirements (blood, platelets, fresh frozen plasma).

  2. Hospital stay.

    1. Total hospital stay.

    2. Intensive therapy unit (ITU) stay.

Search methods for identification of studies

Electronic searches

We searched The Cochrane Central Register of Controlled Trials (CENTRAL) in The Cochrane Library, MEDLINE, EMBASE, and Science Citation Index Expanded (Royle 2003). We also searched the metaRegister of Controlled Trials (mRCT) (http://www.controlled-trials.com/mrct/). The meta-register includes ISRCTN register and NIH ClinicalTrials.gov register among other registers. We have given the search strategies and the time span of the searches in Appendix 1.

Searching other resources

We searched the references of the identified trials to identify further relevant trials.

Data collection and analysis

We performed the systematic review following the instructions given in the Cochrane Handbook for Systematic Reviews of Intervention (Higgins 2011) and the Cochrane Hepato-Biliary Group Module (Gluud 2011).

Selection of studies

Two authors (KG and TP) identified the trials for inclusion independently of each other. We have listed the excluded studies with the reasons for the exclusion. Any differences were resolved through discussion and arbitration by BRD.

Data extraction and management

Both authors independently extracted the following data.

  1. Year and language of publication.

  2. Country.

  3. Year of conduct of the trial.

  4. Inclusion and exclusion criteria.

  5. Sample size.

  6. Donor characteristics such as age, sex, living or cadaveric donor.

  7. Participant characteristics such as age, sex, MELD score, reason for transplant.

  8. Piggy-back or conventional technique: During liver transplantation, the recipient inferior vena cava (IVC) is cross-clamped in the conventional method of liver transplantation. This results in a decrease in venous return from the lower half of the body (Calne 1979). This can result in fall in blood pressure (Calne 1979). To overcome this decreased venous return, methods such as piggy-back technique, use of veno-venous bypass have been attempted. Till date, there is no evidence that these methods have any benefit in the reduction of mortality, morbidity, or blood transfusion requirements (Gurusamy 2011a; Gurusamy 2011b).

  9. Use of veno-venous by-pass (please see above).

  10. Risk of bias (described below).

Any unclear or missing information was sought by contacting the authors of the individual trials. If there was any doubt whether the trials share the same patients - completely or partially (by identifying common authors and centres), we contacted the authors of the trials to clarify whether the trial report has been duplicated. We resolved any differences in opinion through discussion.

Assessment of risk of bias in included studies

We followed the instructions given in the Cochrane Handbook for Systematic Reviews of Intervention (Higgins 2011) and the Cochrane Hepato-Biliary Group Module (Gluud 2011). According to empirical evidence (Schulz 1995; Moher 1998; Kjaergard 2001; Wood 2008), the risk of bias of the trials will be assessed based on the following bias risk domains.

Allocation sequence generation
  • Low risk of bias: sequence generation was achieved using computer random number generation or a random number table. Drawing lots, tossing a coin, shuffling cards, and throwing dice are adequate if performed by an independent adjudicator. 

  • Uncertain risk of bias: the trial is described as randomised, but the method of sequence generation was not specified.

  • High risk of bias: the sequence generation method is not, or may not be, random. Quasi-randomised studies, those using dates, names, or admittance numbers in order to allocate patients are inadequate and will be excluded for the assessment of benefits but not for harms. 

Allocation concealment
  • Low risk of bias: allocation was controlled by a central and independent randomisation unit, sequentially numbered, opaque and sealed envelopes or similar, so that intervention allocations could not have been foreseen in advance of, or during, enrolment. 

  • Uncertain risk of bias: the trial was described as randomised, but the method used to conceal the allocation was not described so that intervention allocations may have been foreseen in advance of, or during, enrolment. 

  • High risk of bias: if the allocation sequence was known to the investigators who assigned participants, or if the study was quasi-randomised. We planned to exclude quasi-randomised studies for the assessment of benefits but not for harms. 

Blinding of participants and personnel
  • Low risk of bias: the trial was described as blinded and the method of blinding of participants and personnel was described so that knowledge of allocation was adequately prevented during the trial.

  • Uncertain risk of bias: the trial was described as blinded, but the method of blinding was not described so that knowledge of allocation was possible during the trial. 

  • High risk of bias, the trial was not blinded so that the allocation was known during the trial.

Blinding of outcome assessors
  • Low risk of bias: the trial was described as blinded and the method of blinding of outcome assessors was described so that knowledge of allocation was adequately prevented during the trial.

  • Uncertain risk of bias: the trial was described as blinded, but the method of blinding was not described so that knowledge of allocation was possible during the trial. 

  • High risk of bias, the trial was not blinded so that the allocation was known during the trial.

Incomplete outcome data
  • Low risk of bias: (the underlying reasons for missingness are unlikely to make treatment effects departure from plausible values, or proper methods have been employed to handle missing data).

  • Uncertain risk of bias: (there is insufficient information to assess whether the missing data mechanism in combination with the method used to handle missing data is likely to induce bias on the estimate of effect).

  • High risk of bias: (the crude estimate of effects (eg, complete case estimate) will clearly be biased due to the underlying reasons for missingness, and the methods used to handle missing data are unsatisfactory).

Selective outcome reporting
  • Low risk of bias: pre-defined, or clinically relevant and reasonably expected outcomes are reported on. 

  • Uncertain risk of bias: not all pre-defined or clinically relevant and reasonably expected outcomes are reported on, or are not reported fully, or it is unclear whether data on these outcomes were recorded or not. 

  • High risk of bias: one or more clinically relevant and reasonably expected outcomes were not reported on; data on these outcomes were likely to have been recorded.

Vested interest bias
  • Low risk of bias: the trial was conducted by parties that might have conflicting interest (eg, drug manufacturer).

  • Uncertain risk of bias: the source of funding was not clear.

  • High risk of bias: the trial was funded by a drug manufacturer.

We considered trials which are classified as low risk of bias in all the above domains as low bias-risk trials.

Measures of treatment effect

For dichotomous variables, we calculated the risk ratio (RR) with 95% confidence interval (CI). Risk ratio calculations do not include trials in which no events occurred in either group, whereas risk difference calculations do. We planned to report the risk difference if the results using this association measure were different from risk ratio. For continuous variables, we calculated the mean difference (MD) with 95% CI for outcomes such as hospital stay and standardised mean difference (SMD) with 95% CI for blood transfusion requirements (which can be reported as units transfused or as ml transfused) and quality of life (where different scales might be used). For time-to-event outcomes such as survival at maximal follow-up, we planned to calculate the hazard ratio (HR) with 95% CI.

Unit of analysis issues

The unit of analysis were the aggregate data on patients who underwent liver transplantation according to randomised group.

Dealing with missing data

We performed an intention-to-treat analysis (Newell 1992) whenever possible. We imputed data for binary outcomes using various scenarios such as best-case, worst-case, best-worst and worse-best scenarios (Gurusamy 2009c; Gluud 2011).

For continuous outcomes, we used available-case analysis. We impute the standard deviation from P values according to the instructions given in the Cochrane Handbook for Systematic Reviews of Intervention (Higgins 2011) and use the median for the meta-analysis when the mean is not available. If it was not possible to calculate the standard deviation from the P value or the confidence intervals, we imputed the standard deviation as the highest standard deviation in the other trials included under that outcome fully recognising that this form of imputation will decrease the weight of the study for calculation of mean differences and bias the effect estimate to no effect in case of standardised mean difference (Higgins 2011).

For time-to-event outcomes, if the hazard ratio and 95% confidence intervals were not reported, we planned to obtain the logarithm of hazard ratios (ln(HR)) and the standard error (SE) of ln(HR) according to the methods described by Parmar 1998 using the Excel sheet provided by Tierney 2007.

Assessment of heterogeneity

We explored heterogeneity by chi-squared test with significance set at P value 0.10, and measure the quantity of heterogeneity by I2 (Higgins 2002).

Assessment of reporting biases

We planned to use visual asymmetry on a funnel plot to explore reporting bias in the presence of at least 10 trials (Egger 1997; Macaskill 2001). We planned to perform linear regression approach described by Egger 1997 to determine the funnel plot asymmetry.

Data synthesis

We performed the meta-analysis using two different approaches. The first approach was the meta-analyses using the software package Review Manager 5 (RevMan 2011) and following the recommendations of The Cochrane Collaboration (Higgins 2011) and the Cochrane Hepato-Biliary Group Module (Gluud 2011). We performed the meta-analysis when there were at least two trials that could be included in the comparison. We used both a random-effects model (DerSimonian 1986) and a fixed-effect model (DeMets 1987) meta-analyses. In case of discrepancy between the two models we have reported both results; otherwise we planned to report the results of the fixed-effect model. We planned to use the generic inverse method to combine the hazard ratios for time-to-event outcomes. If there was only one trial that could be included for the comparison for a binary outcome, we used the Fisher's exact test and present the P value rather than the risk ratio or risk difference. We used the statistical software StatsDirect 2.7 to perform this.

The second approach was network meta-analysis which takes the indirect comparisons into account. We performed the network meta-analysis using SAS 9.2 statistical software. A sample code for a binary outcome and a continuous outcome is shown in Appendix 2. The starting value for the parameters was obtained by running a logistic regression and simple linear regression, which do not take the association of data in different arms at the study level into account.

Subgroup analysis and investigation of heterogeneity

We planned to perform the following subgroup analyses:

  • Trials with low risk of bias compared to trials with high risk of bias.

  • Different methods to decrease blood loss and blood transfusion requirements.

  • Trials using piggy-back technique of liver transplantation versus those using the conventional method of liver transplantation.

  • Trials using veno-venous bypass versus those that do not use veno-venous bypass.

We planned to use the 'test for subgroup differences' to identify the differences between subgroups.

Trial sequential analysis

We used trial sequential analysis (Wetterslev 2008; Wetterslev 2009) using an alpha error of 0.05, beta error of 0.20, control event proportion considered to be 10% based on an approximation of Eurpeon Liver Transplant Registry data (ELTR 2011), and a relative risk reduction of 20% for mortality. The theoretical background and the procedures for trial sequential analysis has been described in detail (Thorlund 2011). We used the trial sequential analysis (TSA) program (TSA 2011).

Sensitivity analysis

We performed a sensitivity analysis by imputing data for binary outcomes using various scenarios such as good outcome analysis, bad outcome analysis, best-case scenario, and worst-case scenario (Gurusamy 2009c; Gluud 2011). For continuous outcomes, we performed a sensitivity analysis by excluding the trials in which the mean and the standard deviation were imputed. We also performed a sensitivity analysis by imputing the standard deviation as the average standard deviation from the trials that reported the outcome in order to assess the impact of the imputation of standard deviation.

Presentation of results

We have presented the results as data tables, forest plots, and as 'Summary of findings' tables in addition to narrative text. We have presented the summary of findings for mortality, primary graft non-function (Summary of findings for the main comparison), thromboembolic episodes and other serious adverse events (Summary of findings 2), blood loss (Summary of findings 3), and red cell or whole blood transfusion (Summary of findings 4).

Results

Description of studies

Results of the search

We identified a total of 3511 bibliographic references through the electronic searches in The Cochrane Central Register of Controlled Trials (CENTRAL) in The Cochrane Library (n = 472), MEDLINE (n = 910), EMBASE (n = 1281), and Science Citation Index Expanded (n = 848). No additional trials were identified through reference search or meta-register of current controlled trials. We did not identify any reference from reference searching. We excluded 1143 duplicates and 2291 clearly irrelevant references through reading abstracts. Seventy-seven references were retrieved for further assessment. We excluded 23 references for the reasons stated in 'Characteristics of excluded studies'. Fifty-four references of 33 completed randomised clinical trials were included in the review.The reference flow is shown in Figure 1.

Figure 1.

Study flow diagram.

Included studies

The characteristics of the patients included in the trials is shown in 'Characteristics of included studies'. All the included trials included patients undergoing liver transplantation. The average age ranged between 43 years and 58 years. The proportion of females ranged between 16% to 55%.

The interventions included pharmacological interventions (aprotinin, tranexamic acid, epsilon amino caproic acid (EACA), antithrombin 3, recombinant factor (rFvIIa), oestrogen, prostaglandin, or epinephrine), blood substitutes (blood components rather than whole blood, hydroxy-ethyl starch), thromboelastography, or cardiovascular interventions (low central venous pressure).

There were 24 two-armed trials. For the negative or inactive controls, 12 trials used placebo (Groh 1993; Ickx 1993; Yassen 1993; Milroy 1995; Boylan 1996; Marcel 1996; Garcia-Huete 1997; Kaspar 1997; Merle 1997; Frenette 1998; Findlay 2001; Pugliese 2007) and nine trials used no intervention (Cottam 1991; Baudo 1992; Himmelreich 1993; Williamson 1999; Hei 2005; Ponnudurai 2005; Wang 2010; Feng 2010; Rummo 2010). Three trials compared one intervention with another (Laine 2003; Dalmau 2004; Ahn 2008). Three trials compared different doses of the same drug (Himmelreich 1992; Soilleux 1995; Lassale 1996). There were six trials with three or more intervention arms. Of these, three trials used different doses of the same drug and compared them with placebo (Porte 2000; Lodge 2005; Planinsic 2005). Three trials used different drugs and compared them with placebo (Ickx 1995; Dalmau 2000; Ickx 2006).

The drug doses used in the studies were highly variable. We calculated the total mean drug dose based on a mean duration of six hours for liver transplantation, using 70 kg patient for those trials in which the drug dose was based on the weight of the patient. The trials were considered to have different levels of doses based on information below. These were arbitrary decisions made after the data extraction was performed but before the data was analysed.

Aprotinin            

  • High dose: > 2.5 million KIU

  • Medium dose: 1 to 2.5 million KIU

  • Low dose: < 1 million KIU

Tranexamic acid and recombinant factor VIIa (rFVIIa)                  

  • High dose: > =100 mg/Kg

  • Low dose: < 100 mg/Kg

The doses in the individual trial are shown in Table 1.

Table 1. Doses used in the trials
  1. * Control was placebo or no intervention unless stated.
    ITU = intensive therapy unit.

 BolusContinuous/hourAdditional bolusesStart timeEnd timeApproximate total including bolus
Aprotinin (million KIU units)
Aprotinin versus control*
Cottam 199120.50.05 per unit transfusedinductionnot stated5
Findlay 200110.25Nonenot statedend of surgery2.5
Dalmau 2004 (control: tranexamic acid)20.5NoneBeginning of surgery2 hours after reperfusion5
Garcia-Huete 199720.5Noneinductionend of surgery5
Groh 199320.5Noneinductionend of surgery5
Hei 2005None0.4Noneinductionnot stated2.4
Ickx 1993None0.2NoneBeginning of surgerynot stated1.2
Ickx 199520.5Noneanhepatic phasenot stated5
Ickx 2006 (control: tranexamic acid)20.5Noneanhepatic phase2 hours after reperfusion3.5
Marcel 1996None0.2Noneinductionnot stated1.2
Merle 199720.5NoneBeginning of surgeryend of surgery5
Milroy 199520.5Noneinductiontransfer to ITU5
Porte 2000 (high dose A)211 million KIU half an hour before reperfusioninduction2 hours after reperfusion8
Porte 2000 (high dose B)20.5Noneinduction2 hours after reperfusion5.5
Different doses or methods of administration of aprotinin
Himmelreich 1992 (bolus method)0.5None0.5 during anhepatic phase and then at reperfusioninductionreperfusion1
Himmelreich 1992 (continuous method)None0.2, increased to 0.4 during anhepatic phase and decreased to 0.1 from reperfusionNoneinductionend of surgery1.2
Lassale 1996 (high dose)20.5Noneinductionend of surgery5
Lassale 1996 (low dose)None0.1Noneinductionend of surgery0.6
Soilleux 1995 (high dose)20.5Noneinductiontransfer to ITU5
Soilleux 1995 (medium dose)0.50.15Noneinductiontransfer to ITU1.4

Tranexamic acid versus control*

Units: mg/kg/hr

Boylan 1996None40Noneinductionreperfusion240
Dalmau 2000None10Noneinductionreperfusion60
Dalmau 2004 (control: aprotinin)None10NoneBeginning of surgery2 hours after reperfusion60
Ickx 19958040Noneanhepatic phasenot stated320
Ickx 2006 (control: aprotinin)4040Noneanhepatic phase2 hours after reperfusion280
Kaspar 1997None2NoneBeginning of surgeryend of surgery12
Yassen 1993103Noneanhepatic phasetransfer to ITU28

Epsilon amino caproic acid (EACA) versus control*

Units: mg/kg/hour

Dalmau 2000None16Noneinductionreperfusion96

Antithrombin versus control*

Units: units per hour

Baudo 1992(100-plasma activity level) per kg body weight1000Noneinduction end of surgery6000

Recombinant Factor VIIa versus control*

(mcg/kg)

Lodge 2005 (high dose A)60None60 mcg/kg every 2 hours to approximately 30 minutes before reperfusion and a final dose on wound closurewithin 10 minutes of skin incisionNone180
Lodge 2005 (high dose B)120None120 mcg/kg every 2 hours to approximately 30 minutes before reperfusion and a final dose on wound closurewithin 10 minutes of skin incisionNone360
Planinsic 2005 (low dose A)20NoneNonewithin 10 minutes of skin incision (single bolus)None20
Planinsic 2005 (low dose B)40NoneNonewithin 10 minutes of skin incision (single bolus)None40
Planinsic 2005 (low dose C)80NoneNonewithin 10 minutes of skin incision (single bolus)None80
Pugliese 200740Just before inductionNoneJust before inductionNone40

Oestrogen versus control*

Units: mg

Frenette 1998100None100 mg 30 min after reperfusionBeginning of surgeryNone200

Prostaglandin E versus control*

Units: microgram (mcg)

Himmelreich 1993None10 mcg increased to 40 mcgNoneBeginning of surgerythree post-operative days2800

Norpeinephrine versus control*

Units: mcg/min

Ponnudurai 2005None0.5 mcg/min increased by 1 mcg increments up to a maximum of 6 mcg/min to maintain a systolic blood pressure and pulmonary capillary wedge pressure more than 80% of baseline values.

Risk of bias in included studies

The risk of bias is summarised in Figure 2 and Figure 3.

Figure 2.

Risk of bias graph: review authors' judgements about each risk of bias item presented as percentages across all included studies.

Figure 3.

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

Aprotinin versus control

Eight trials were included under this comparison (Groh 1993; Ickx 1993; Marcel 1996; Garcia-Huete 1997; Merle 1997; Porte 2000; Findlay 2001; Hei 2005). All the trials were of high risk of bias.

Tranexamic acid versus control

Five trials were included under this comparison (Yassen 1993; Boylan 1996; Kaspar 1997; Dalmau 2000; Pugliese 2007). All the trials were of high risk of bias.

Epsilon amino caproic acid versus control

Only one trial was included under this comparison (Dalmau 2000). This trial was of high risk of bias.

Antithrombin III versus control

Only one trial was included under this comparison (Baudo 1992). This trial was of high risk of bias.

Recombinant factor VIIa versus control

Three trials were included under this comparison (Lodge 2005; Planinsic 2005; Pugliese 2007). All the trials were of high risk of bias.

Oestrogen versus control

Only one trial was included under this comparison (Frenette 1998). This trial was of high risk of bias.

Prostaglandin versus control

Only one trial was included under this comparison (Himmelreich 1993). This trial was of high risk of bias.

Norepinephrine versus control

Only one trial was included under this comparison (Ponnudurai 2005). This trial was of high risk of bias.

Thromboelastography versus control

Two trials were included under this comparison (Rummo 2010; Wang 2010). Both the trials were of high risk of bias.

Low central venous pressure (CVP) versus control

Only one trial was included under this comparison (Feng 2010). This trial was of high risk of bias.

Aprotinin: bolus versus continuous infusion

Only one trial was included under this comparison (Himmelreich 1992). This trial was of high risk of bias.

Aprotinin: high dose versus medium dose

Only one trial was included under this comparison (Soilleux 1995). This trial was of high risk of bias.

Aprotinin: high dose versus low dose

Only one trial was included under this comparison (Lassale 1996). This trial was of high risk of bias.

Whole blood versus blood components

Only one trial was included under this comparison (Laine 2003). This trial was of high risk of bias.

Solvent detergent plasma versus standard fresh frozen plasma

Only one trial was included under this comparison (Williamson 1999). This trial was of high risk of bias.

Tranexamic acid versus epsilon amino caproic acid

Only one trial was included under this comparison (Dalmau 2000). This trial was of high risk of bias.

Tranexamic acid versus aprotinin

Three trials were included under this comparison (Ickx 1995; Dalmau 2004; Ickx 2006). All the trials were of high risk of bias.The trials were included for the outcomes that they reported.

Effects of interventions

See: Summary of findings for the main comparison Intervention versus control for liver transplantation (mortality, primary graft non-function, and retransplantation); Summary of findings 2 Intervention versus control for liver transplantation (thromboembolic episodes and other serious adverse events); Summary of findings 3 Intervention versus control for liver transplantation (blood loss); Summary of findings 4 Intervention versus control for liver transplantation (red cell or whole blood allogeneic transfusion)

Interventions versus inactive controls

Aprotinin versus control

Eight trials were included under this comparison (Groh 1993; Ickx 1993; Marcel 1996; Garcia-Huete 1997; Merle 1997; Porte 2000; Findlay 2001; Hei 2005). All the trials were of high risk of bias.The trials were included for the outcomes that they reported.

Primary outcomes

Mortality

There was no significant difference in the 60-day mortality between the aprotinin groups versus the control (RR 0.52; 95% CI 0.18 to 1.45) (Analysis 1.1).

Graft failure and retransplantation
There was no significant difference in the primary graft non-function (RR 0.15; 95% CI 0.02 to 1.25) (Analysis 1.3). Long-term graft failure was not reported in any of the trials. The proportion of patients undergoing retransplantation was significantly lower in patients receiving aprotinin compared to placebo by the fixed-effect model (RR 0.18; 95% CI 0.03 to 0.90) (Analysis 1.4) but not by random-effects model (RR 0.21; 95% CI 0.02 to 1.79). There was no significant statistical heterogeneity (I2 = 26%; P = 0.25).

Serious adverse events
There was no significant difference in the proportion of patients who developed thromboembolic episodes (RR 0.59; 95% CI 0.19 to 1.84) (Analysis 1.5) or other serious adverse events between the groups (Fisher's exact test: P > 0.99).

Quality of life
This outcome was not reported in any of the trials.

Secondary outcomes  

Blood loss and allogenic transfusion requirements
There was no significant difference in the blood loss between the two groups (MD -1.36 litre; 95% CI -3.39 to 0.66) (Analysis 1.7). There was no significant difference in the proportion of patients requiring allogeneic red cell blood transfusion (Fisher's exact test: P = 0.18). There was significantly lower allogeneic blood transfusion requirements, amount of platelets, fresh frozen plasma, and cryoprecipitate transfused in the aprotinin group than the control group (SMD -0.59; 95% CI -0.80 to -0.39 (Analysis 1.9); SMD -0.44; 95% CI -0.73 to -0.14 (Analysis 1.10); SMD -0.33; 95% CI -0.53 to -0.13 (Analysis 1.11); SMD -0.49; 95% CI -0.82 to -0.16 (Analysis 1.12) respectively).

Hospital stay
There was no significant difference in the mean total hospital stay or ITU stay between the two groups (MD 0.00 days; 95% CI -4.94 to 4.94 (Analysis 1.13); MD -0.21 days; 95% CI -0.77 to 0.36 (Analysis 1.14) respectively).

Tranexamic acid versus control

Five trials were included under this comparison (Yassen 1993; Boylan 1996; Kaspar 1997; Dalmau 2000; Pugliese 2007). All the trials were of high risk of bias.The trials were included for the outcomes that they reported.

Primary outcomes

Mortality
There was no significant difference in the 60-day mortality between the tranexamic acid versus control groups (RR 0.55; 95% CI 0.17 to 1.76) (Analysis 1.1).

Graft failure and retransplantation
There was no significant difference in the primary graft non-function (Fisher's exact test: P > 0.99) (Kaspar 1997). Long-term graft failure was not reported in any of the trials. There was no significant difference in the proportion of patients undergoing retransplantation (RR 0.79; 95% CI 0.18 to 3.48) (Analysis 1.4).

Serious adverse events
There was no significant difference in the proportion of patients who developed thromboembolic episodes (RR 2.20; 95% CI 0.38 to 12.64) (Analysis 1.5). None of the trials reported other serious adverse events.

Quality of life
This outcome was not reported in any of the trials.

Secondary outcomes  

Blood loss and allogenic transfusion requirements

There was no significant difference in the blood loss between the two groups (MD -4.98 litre; 95% CI -10.18 to 0.23) (Analysis 1.7). There was no significant difference in the proportion of patients requiring allogeneic red cell blood transfusion (Fisher's exact test: P = 0.11) (Dalmau 2000). There was no significant difference in the allogeneic blood transfusion requirements, amount of platelets, fresh frozen plasma, or cryoprecipitate transfused (SMD -0.27; 95% CI -0.59 to 0.06 (Analysis 1.9); SMD -0.17; 95% CI -0.49 to 0.16 (Analysis 1.10); SMD -0.21; 95% CI -0.53 to 0.11 (Analysis 1.11); SMD -0.27; 95% CI -0.57 to 0.02 (Analysis 1.12) respectively) between the two groups.

Hospital stay

There was no significant difference in the mean total hospital stay or ITU stay between the two groups (MD 4.00 days; 95% CI -16.18 to 24.18 (Analysis 1.13); MD -1.00 days; 95% CI -4.75 to 2.75 (Analysis 1.14) respectively).

Epsilon amino caproic acid versus control

Only one trial was included under this comparison (Dalmau 2000). This trial was of high risk of bias.

Primary outcomes

Mortality

There was no significant difference in the 60-day mortality between the groups (Fisher's exact test: P = 0.65).

Graft failure and retransplantation

None of these outcomes were reported in this trial.

Serious adverse events

There was no significant difference in the proportion of patients who developed thromboembolic episodes (Fisher's exact test: P > 0.99). Other serious adverse events were not reported in this trial.

Quality of life

This outcome was not reported in any of the trials.

Secondary outcomes  

Blood loss and allogenic transfusion requirements

The amount of blood loss was not reported in this trial. There was no significant difference in the proportion of patients requiring allogeneic red cell blood transfusion (Fisher's exact test: P = 0.41). There was no significant difference in the allogeneic blood transfusion requirements, amount of platelets, fresh frozen plasma, or cryoprecipitate transfused (SMD -0.17; 95% CI -0.71 to 0.36 (Analysis 1.9); SMD 0.01; 95% CI -0.52 to 0.55 (Analysis 1.10); SMD -0.12; 95% CI -0.66 to 0.41 (Analysis 1.11); SMD 0.00; 95% CI -0.53 to 0.54 (Analysis 1.12) respectively) between the two groups.

Hospital stay

The mean total hospital stay or ITU stay were not reported by this trial.

Antithrombin III versus control

Only one trial was included under this comparison (Baudo 1992). This trial was of high risk of bias.

Primary outcomes

Mortality

There was no significant difference in the 60-day mortality or mortality at maximal follow-up between the antithrombin III versus the control groups (Fisher's exact test: P = 0.49 and P = 0.18 respectively).

Graft failure and retransplantation

None of these outcomes were reported in this trial.

Serious adverse events

The proportion of patients who developed thromboembolic episodes or other serious adverse events were not reported in this trial.

Quality of life

This outcome was not reported in this trial.

Secondary outcomes  

Blood loss and allogenic transfusion requirements

There was no significant difference in the amount of blood loss between the two groups (MD 1.90 litre; 95% CI -4.86 to 8.66) (Analysis 1.7). The proportion of patients requiring allogeneic red cell blood transfusion was not reported in this trial. There was no significant difference in the allogeneic blood transfusion requirements, amount of platelets, fresh frozen plasma, or cryoprecipitate transfused (SMD -0.04; 95% CI -0.77 to 0.69 (Analysis 1.9); SMD 0.00; 95% CI -0.73 to 0.73 (Analysis 1.10); SMD 0.04; 95% CI -0.69 to 0.77 (Analysis 1.11); SMD 0.00; 95% CI -0.73 to 0.73 (Analysis 1.12) respectively) between the two groups.

Hospital stay

The mean total hospital stay or ITU stay were not reported in this trial.

Recombinant factor VIIa versus control

Three trials were included under this comparison (Lodge 2005; Planinsic 2005; Pugliese 2007). All the trials were of high risk of bias.

Primary outcomes

Mortality

There was no significant difference in the 60-day mortality between the recombinant factor VIIa versus the control groups (RR 1.51; 95% CI 0.33 to 6.95).

Graft failure and retransplantation

There were no primary graft non-function in the only trial that reported the outcome (Pugliese 2007). There was no significant difference in the graft failure (Fisher's exact test: P > 0.99). The proportion of patients who underwent retransplantation was not reported in any of the two trials.

Serious adverse events

There was no significant difference in the proportion of patients who developed thromboembolic episodes (RR 1.38; 95% CI 0.65 to 2.91) (Analysis 1.5) or other serious adverse events between the groups (RR 1.30; 95% CI 0.94 to 1.78) (Analysis 1.6).

Quality of life

This outcome was not reported in any of the trials.

Secondary outcomes  

Blood loss and allogenic transfusion requirements

The amount of blood loss was not reported in the two trials. The proportion of patients requiring allogeneic red cell blood transfusion was lower in the rVIIIa group (Fisher's exact test: P = 0.02). There was no significant difference in the allogeneic blood transfusion requirements, amount of platelets, or fresh frozen plasma transfused (SMD -0.05; 95% CI -0.32 to 0.23 (Analysis 1.9); SMD -0.03; 95% CI -0.31 to 0.25 (Analysis 1.10); SMD -0.01; 95% CI -0.28 to 0.25 (Analysis 1.11) respectively) between the two groups.

Hospital stay

There was no significant difference in the mean total hospital stay or ITU stay between the two groups (MD 3.40 days; 95% CI -7.51 to 14.31(Analysis 1.13); MD 0.22 days; 95% CI -0.14 to 0.58 (Analysis 1.14) respectively).

Oestrogen versus control

Only one trial was included under this comparison (Frenette 1998). This trial was of high risk of bias.

Primary outcomes

Mortality

This outcome was not reported in this trial.

Graft failure and retransplantation

None of these outcomes were reported in this trial.

Serious adverse events

The proportion of patients who developed thromboembolic episodes or other serious adverse events were not reported in this trial.

Quality of life

This outcome was not reported in this trial.

Secondary outcomes  

Blood loss and allogenic transfusion requirements

The amount of blood loss and the proportion of patients requiring allogeneic red cell blood transfusion were not reported in this trial. There was no significant difference in the allogeneic blood transfusion requirements, amount of platelets, fresh frozen plasma, or cryoprecipitate transfused (SMD -0.58; 95% CI -1.32 to 0.15 (Analysis 1.9); SMD -0.53; 95% CI -1.26 to 0.20 (Analysis 1.10); SMD -0.67; 95% CI -1.41 to 0.07 (Analysis 1.11); SMD 0.00; 95% CI -0.72 to 0.72 (Analysis 1.12) respectively) between the two groups.

Hospital stay

The mean total hospital stay or ITU stay were not reported by this trial.

Prostaglandin versus control

Only one trial was included under this comparison (Himmelreich 1993). This trial was of high risk of bias.

Primary outcomes

Mortality

This outcome was not reported in this trial.

Graft failure and retransplantation

Primary graft function and long-term graft failure were not reported in this trial. There were no retransplantations in either group in this trial.

Serious adverse events

The proportion of patients who developed thromboembolic episodes or other serious adverse events were not reported in this trial.

Quality of life

This outcome was not reported in this trial.

Secondary outcomes  

Blood loss and allogenic transfusion requirements

The amount of blood loss and the proportion of patients requiring allogeneic red cell blood transfusion were not reported in this trial. There was no significant difference in the allogeneic blood transfusion requirements or amount of cryoprecipitate transfused (SMD -0.04; 95% CI -0.92 to 0.83 (Analysis 1.9); SMD 0.00; 95% CI -0.88 to 0.88 (Analysis 1.11) respectively) between the two groups.

Hospital stay

The mean total hospital stay or ITU stay were not reported by this trial.

Norepinephrine versus control

Only one trial was included under this comparison (Ponnudurai 2005). This trial was of high risk of bias.

Primary outcomes

Mortality

This outcome was not reported in this trial.

Graft failure and retransplantation

None of these outcomes were reported in this trial.

Serious adverse events

The proportion of patients who developed thromboembolic episodes or other serious adverse events were not reported in this trial.

Quality of life

This outcome was not reported in this trial.

Secondary outcomes  

Blood loss and allogenic transfusion requirements

The amount of blood loss and the proportion of patients requiring allogeneic red cell blood transfusion were not reported in this trial. There was no significant difference in the allogeneic blood transfusion requirements, amount of platelets, or fresh frozen plasma transfused between the norepinephrine versus the control group (SMD -0.04; 95% CI -0.53 to 0.44 (Analysis 1.9); SMD 0.34; 95% CI -0.15 to 0.83 (Analysis 1.10); SMD 0.11; 95% CI -0.38 to 0.60 (Analysis 1.11) respectively) between the two groups.

Hospital stay

There was no significant difference in the mean total hospital stay or ITU stay between the two groups (MD -1.00 days; 95% CI -3.18 to 1.18 (Analysis 1.13); MD -0.13 days; 95% CI -1.65 to 1.39 (Analysis 1.14) respectively).

Thromboelastography versus control

Two trials were included under this comparison (Rummo 2010; Wang 2010). Both the trials were of high risk of bias.

Primary outcomes

Mortality

There was no significant difference in the mortality at maximal follow-up between the thromboelastography versus the control groups (Fisher's exact test: P > 0.99) (Wang 2010).

Graft failure and retransplantation

None of these outcomes were reported in either trial.

Serious adverse events

The proportion of patients who developed thromboembolic episodes or other serious adverse events were not reported in either trial.

Quality of life

This outcome was not reported in either trial.

Secondary outcomes  

Blood loss and allogenic transfusion requirements
There was no significant difference in the amount of blood loss between the two groups (SMD -0.73; 95% CI -1.69 to 0.24) (Analysis 1.7). The proportion of patients requiring allogeneic red cell blood transfusion was not reported in either trial. The allogeneic blood transfusion requirements were significantly lower in the thromboelastography group than controls by the fixed effect model (SMD -0.73; 95% CI -1.25 to -0.20) but not by the random-effects model (SMD -0.73; 95% CI -1.69 to 0.24) (Analysis 1.9). There was significant statistical heterogeneity (I2 = 70%; P = 0.07). There was no significant difference in the amount of platelet transfusion, or cryoprecipitate (SMD -0.17; 95% CI -0.91 to 0.58 (Analysis 1.10) ; SMD -0.25; 95% CI -1.00 to 0.49 (Analysis 1.12) respectively) between the two groups.

The amount of fresh frozen plasma transfused was significantly lower in the thromboelastography group than control (SMD -0.82; 95% CI -1.60 to -0.05) (Analysis 1.11).

Hospital stay

The mean total hospital stay or ITU stay were not reported in this trial.

Low central venous pressure (CVP) versus control

Only one trial was included under this comparison (Feng 2010). This trial was of high risk of bias.

Primary outcomes

Mortality

This outcome was not reported in this trial.

Graft failure and retransplantation

None of these outcomes were reported in this trial.

Serious adverse events
The proportion of patients who developed thromboembolic episodes was not reported in this trial. The proportion of patients with other serious adverse events was significantly lower in the low CVP group than the control group (Fisher's exact test: P = 0.01)

Quality of life

This outcome was not reported in this trial.

Secondary outcomes  

Blood loss and allogenic transfusion requirements

The amount of blood loss was significantly lower in the low CVP group than control (MD -1.19 litre; 95% CI -1.88 to -0.50) (Analysis 1.7). The proportion of patients requiring allogeneic red cell blood transfusion was not reported in this trial. The allogeneic blood transfusion requirements were significantly lower in the low CVP group than the control group (SMD -1.17; 95% CI -1.62 to -0.71) (Analysis 1.9).

Hospital stay

There was no significant difference in the mean total hospital stay or ITU stay between the two groups (MD 0.00 days; 95% CI -19.43 to 19.43 (Analysis 1.13); (MD -1.00 days; 95% CI -3.71 to 1.71 (Analysis 1.14)).

One intervention versus another intervention

Aprotinin: bolus versus continuous infusion

Only one trial was included under this comparison (Himmelreich 1992). This trial was of high risk of bias.

Primary outcomes

Mortality

There was no mortality in either group.

Graft failure and retransplantation

Primary graft non-function and long-term graft failure were not reported in this trial. There was no significant difference in the proportion of patients requiring retransplantation between the two groups (Fisher's exact test: P > 0.99).

Serious adverse events
The proportion of patients who developed thromboembolic episodes or other serious adverse events were not reported in this trial.

Quality of life

This outcome was not reported in this trial.

Secondary outcomes  

Blood loss and allogenic transfusion requirements
The amount of blood loss and the proportion of patients requiring allogeneic red cell blood transfusion were not reported in this trial. There was no significant difference in the allogeneic blood transfusion requirements or the amount of fresh frozen plasma transfused (SMD 0.86; 95% CI -0.00 to 1.73 (Analysis 1.9); SMD 0.05; 95% CI -0.78 to 0.87 (Analysis 1.11) respectively) between the two groups.

Hospital stay

The mean total hospital stay was not reported in this trial. The ITU stay was significantly lower in the bolus group than the continuous infusion group (MD 2.00 days; 95% CI -3.99 to 7.99 (Analysis 1.14) respectively).

Aprotinin: high dose versus medium dose

Only one trial was included under this comparison (Soilleux 1995). This trial was of high risk of bias.

Primary outcomes

Mortality
This outcome was not reported in this trial.

Graft failure and retransplantation
None of these outcomes were reported in this trial.

Serious adverse events
The proportion of patients who developed thromboembolic episodes or other serious adverse events were not reported in this trial.

Quality of life
This outcome was not reported in this trial.

Secondary outcomes  

Blood loss and allogenic transfusion requirements
The amount of blood loss and the proportion of patients requiring allogeneic red cell blood transfusion were not reported in this trial. There was no significant difference in the allogeneic blood transfusion requirements, amount of platelets, or fresh frozen plasma transfused (SMD 0.16; 95% CI -0.12 to 0.45 (Analysis 1.9); SMD -0.48; 95% CI -0.77 to -0.19 (Analysis 1.10); SMD 0.06; 95% CI -0.22 to 0.35 (Analysis 1.11) respectively) between the two groups.

Hospital stay
The mean total hospital stay or ITU stay were not reported in this trial.

Aprotinin: high dose versus low dose

Only one trial was included under this comparison (Lassale 1996). This trial was of high risk of bias.

Primary outcomes

Mortality
This outcome was not reported in this trial.

Graft failure and retransplantation
None of these outcomes were reported in this trial.

Serious adverse events
The proportion of patients who developed thromboembolic episodes or other serious adverse events were not reported in this trial.

Quality of life
This outcome was not reported in this trial.

Secondary outcomes  

Blood loss and allogenic transfusion requirements
The amount of blood loss and the proportion of patients requiring allogeneic red cell blood transfusion were not reported in this trial. There was no significant difference in the allogeneic blood transfusion requirements or amount of fresh frozen plasma transfused (SMD -1.46; 95% CI -2.47 to -0.45 (Analysis 1.9); SMD -1.31; 95% CI -2.29 to -0.32 (Analysis 1.11) respectively) between the two groups.

Hospital stay
The mean total hospital stay or ITU stay were not reported in this trial.

Whole blood versus blood components

Only one trial was included under this comparison (Laine 2003). This trial was of high risk of bias.

Primary outcomes

Mortality
This outcome was not reported in this trial.

Graft failure and retransplantation
None of these outcomes were reported in this trial.

Serious adverse events
The proportion of patients who developed thromboembolic episodes or other serious adverse events were not reported in this trial.

Quality of life
This outcome was not reported in this trial.

Secondary outcomes  

Blood loss and allogenic transfusion requirements
The amount of blood loss and the proportion of patients requiring allogeneic red cell blood transfusion were not reported in this trial. There was no significant difference in the allogeneic blood transfusion requirements or the amount of platelets, fresh frozen plasma, or cryoprecipitate transfused (SMD -0.04; 95% CI -0.72 to 0.65 (Analysis 1.9); SMD -0.12; 95% CI -0.81 to 0.56 (Analysis 1.10); SMD -1.53; 95% CI -2.32 to -0.74 (Analysis 1.11); SMD -0.28; 95% CI -0.97 to 0.41 (Analysis 1.12) respectively) between the two groups.

Hospital stay
The mean total hospital stay or ITU stay were not reported in this trial.

Solvent detergent plasma versus standard fresh frozen plasma

Only one trial was included under this comparison (Williamson 1999). This trial was of high risk of bias.

Primary outcomes

Mortality
There was no 60-day mortality in either group in this trial.

Graft failure and retransplantation
None of these outcomes were reported in this trial.

Serious adverse events
The proportion of patients who developed thromboembolic episodes or other serious adverse events were not reported in this trial.

Quality of life
This outcome was not reported in this trial.

Secondary outcomes  

Blood loss and allogenic transfusion requirements: The amount of blood loss and the proportion of patients requiring allogeneic red cell blood transfusion were not reported in this trial. There was no significant difference in the allogeneic blood transfusion requirements or the amount of platelets, fresh frozen plasma, or cryoprecipitate transfused (SMD 0.19; 95% CI -0.60 to 0.98 (Analysis 1.9); SMD 0.07; 95% CI -0.71 to 0.86 (Analysis 1.10); SMD 0.10; 95% CI -0.68 to 0.89 (Analysis 1.12) respectively) between the two groups.

Hospital stay: The mean total hospital stay or ITU stay were not reported in this trial.

Tranexamic acid versus epsilon amino caproic acid

Only one trial was included under this comparison (Dalmau 2000). This trial was of high risk of bias.

Primary outcomes

Mortality
There was no significant difference in the 60-day mortality between the groups (Fisher's exact test: P > 0.99).

Graft failure and retransplantation
None of these outcomes were reported in this trial.

Serious adverse events
There was no significant difference in the proportion of patients who developed thromboembolic episodes (Fisher's exact test: P =0.68). Other serious adverse events were not reported in this trial.

Quality of life

This outcome was not reported in any of the trials.

Secondary outcomes  

Blood loss and allogenic transfusion requirements
The amount of blood loss was not reported in this trial. There was no significant difference in the proportion of patients requiring allogeneic red cell blood transfusion (Fisher's exact test: P = 0.12). There was no significant difference in the allogeneic blood transfusion requirements, amount of platelet transfusion, fresh frozen plasma, or cryoprecipitate (SMD -0.23; 95% CI -0.66 to 0.20 (Analysis 1.9); SMD 0.05; 95% CI -0.37 to 0.48 (Analysis 1.10); SMD -0.22; 95% CI -0.65 to 0.21 (Analysis 1.11); SMD -0.15; 95% CI -0.69 to 0.38) (Analysis 1.12) respectively) between the two groups.

Hospital stay
The mean total hospital stay or ITU stay were not reported by this trial.

Tranexamic acid versus aprotinin

Three trials were included under this comparison (Ickx 1995; Dalmau 2004; Ickx 2006). All the trials were of high risk of bias.The trials were included for the outcomes that they reported.

Primary outcomes

Mortality
There was no significant difference in the 60-day mortality between the groups (RR 4.12; 95% CI 0.71 to 23.76) (Analysis 1.1).

Graft failure and retransplantation
Primary graft non-function and graft failure were not reported in any of the trials. There was no significant difference in the proportion of patients requiring retransplantation between the two groups (Fisher's exact test: P >0.99) (Dalmau 2004).

Serious adverse events
There was no significant difference in the proportion of patients who developed thromboembolic episodes (Fisher's exact test: P =0.68) (Dalmau 2004). Other serious adverse events were not reported in this trial.

Quality of life

This outcome was not reported in any of the trials.

Secondary outcomes  

Blood loss and allogenic transfusion requirements
There was no significant difference in the amount of blood loss between the two groups (MD -1.01 litre; 95% CI -2.31 to 0.29). There was no significant difference in the proportion of patients requiring allogeneic red cell blood transfusion (Fisher's exact test: P = 0.72) (Dalmau 2004). There was no significant difference in the allogeneic blood transfusion requirements, amount of platelets, or fresh frozen plasma transfused (SMD -0.09; 95% CI -0.36 to 0.19 (Analysis 1.9); (SMD -0.08; 95% CI -0.36 to 0.20) (Analysis 1.10); (SMD 0.02; 95% CI -0.45 to 0.48) (Analysis 1.11) respectively) between the two groups.

Hospital stay
There was no significant difference in the mean total hospital stay or ITU stay between the two groups (MD 5.00 days; 95% CI -2.69 to 12.69 (Analysis 1.13); MD 0.00 days; 95% CI -2.40 to 2.40) (Analysis 1.14).

Network meta-analysis

The results of the network meta-analysis are summarised in Table 2.

Table 2. Network analysis results
  1. CVP = central venous pressure.
    EACA = epsilon amino caproic acid.
    FFP = fresh frozen plasma.
    rFVIIa = recombinant factor VIIa.

 P valueEffect estimate
Point estimate and 95% confidence intervals (CI)
60-day mortality
Solvent detergent FFP0.5483RR 0; 95% CI 0 to 257517954.02
Aprotinin0.0489RR 0.09; 95% CI 0.01 to 0.99
Tranexamic acid0.0237RR 0.44; 95% CI 0.22 to 0.88
EACA0.387RR 0.45; 95% CI 0.06 to 3.16
Antithrombin III0.546RR 0; 95% CI 0 to 225447979.34
rFVIIa0.4494RR 0.31; 95% CI 0.01 to 8.19
Mortality at maximal follow-up
Not performed
Primary graft non-function
Convergence not obtained
Graft failure
Not performed
Retransplantation
Convergence not obtained
Thromboembolic episodes
rFVIIa0.0842RR 2.23; 95% CI 0.88 to 5.69
control0.2142RR 0.5; 95% CI 0.16 to 1.6
Aprotinin0.9371RR 1.04; 95% CI 0.36 to 2.99
Tranexamic acid0.8742RR 0.88; 95% CI 0.14 to 5.46
Serious adverse events
Low CVP0.1552RR 0.29; 95% CI 0.03 to 3.13
Aprotinin0.8512RR 0.91; 95% CI 0.13 to 6.24
rFVIIa0.2517RR 1.54; 95% CI 0.48 to 4.88
Blood loss
Aprotinin0.4028MD 0.3 litre; 95% CI - 0.47 to 1.07
Tranexamic acid0.6075MD 0.20 litre; 95% CI -0.63 to 1.03
Antithrombin III0.9966MD 0.03 litre; 95% CI -14.17 to 14.22
Thromboelastometry0.0003MD -1.83 litre; 95% CI -2.55 to -1.12
Low CVP0.004MD -1.12 litre; 95% CI -1.19 to -0.49
Proportion of patients requiring red cell or whole blood transfusion
rFVIIa0.2765RR 0.18; 95% CI 0 to 25.84
Aprotinin0.1301RR 0.37; 95% CI 0.07 to 2.04
Tranexamic acid0.0937RR 0.27; 95% CI 0.04 to 1.72
EACA0.5714RR 0.67; 95% CI 0.05 to 8.95
Amount of blood transfused
Convergence not obtained
Platelet transfusion
Aprotinin0.0633SMD -0.25; 95% CI -0.51 to 0.02
Tranexamic acid0.0129SMD -0.33; 95% CI -0.58 to -0.08
EACA0.0231SMD -0.5; 95% CI -0.92 to -0.08
rFVIIa0.7865SMD -0.04; 95% CI -0.32 to 0.25
Oestrogen0.438SMD 0.28; 95% CI -0.46 to 1.02
Thromboelastometry0.1046SMD 0.62; 95% CI -0.14 to 1.38
Norepinephrine0.1552SMD 0.36; 95% CI -0.15 to 0.87
Solvent detergent FFP0.3926SMD 0.34; 95% CI -0.48 to 1.17
Fresh frozen plasma
Aprotinin0.4663SMD -0.12; 95% CI -0.46 to 0.22
Tranexamic acid0.7514SMD -0.06; 95% CI -0.47 to 0.35
EACA0.9742SMD 0.01; 95% CI -0.56 to 0.58
Antithrombin III0.9217SMD -0.05; 95% CI -1.17 to 1.07
rFVIIa0.9748SMD 0.01; 95% CI -0.64 to 0.66
Oestrogen0.5834SMD -0.3; 95% CI -1.4 to 0.81
Prostaglandin0.8881SMD -0.08; 95% CI -1.31 to 1.14
Thromboelastometry0.7947SMD 0.14; 95% CI -0.98 to 1.27
Norepinephrine0.628SMD -0.23; 95% CI -1.18 to 0.73
Cryoprecipate
Aprotinin0.2675SMD -0.32; 95% CI -0.94 to 0.3
Tranexamic acid0.2572SMD -0.35; 95% CI -1.01 to 0.31
EACA0.5333SMD -0.29; 95% CI -1.32 to 0.74
Oestrogen0.3963SMD 0.36; 95% CI -0.56 to 1.27
Thromboelastometry0.8342SMD 0.1; 95% CI -0.92 to 1.11
Solvent detergent FFP0.4925SMD 0.31; 95% CI -0.68 to 1.29
Hospital stay
Aprotinin0.8137MD 0.59 days; 95% CI -5.89 to 7.06
Tranexamic acid0.1876MD 6.45 days; 95% CI -4.83 to 17.74
rFVIIa0.5458MD 3.04 days; 95% CI -9.78 to 15.87
Low CVP0.1629MD 11.91 days; 95% CI -7.46 to 31.28
Norepinephrine0.3781MD -1.1 days; 95% CI -4.19 to 1.99
Intensive therapy unit stay
Aprotinin0.723MD -0.11 days; 95% CI -0.8 to 0.59
Tranexamic acid0.9696MD 0.04 days; 95% CI -2.35 to 2.43
rFVIIa0.6915MD -0.21 days; 95% CI -1.45 to 1.03
Low CVP0.6431MD 0.6 days; 95% CI -2.39 to 3.59
Norepinephrine0.8102MD 0.19 days; 95% CI -1.67 to 2.06
Primary outcomes

Mortality
The 60-day mortality was significantly lower in the aprotinin (RR 0.09; 95% CI 0.01 to 0.99) and tranexamic acid (RR 0.44; 95% CI 0.22 to 0.88) groups compared with no intervention or placebo. We did not perform the network meta-analysis for mortality at maximal follow-up because of the few trials included under this outcome.

Graft failure and retransplantation
Convergence was not obtained for primary graft non-function and retransplantation. We did not perform the network meta-analysis for graft failure since only one trial reported this outcome.

Serious adverse events
The network meta-analysis did not show any significant differences in thromboembolic episodes or in serious adverse events in the different interventions.

Quality of life
None of the trials reported the quality of life in patients.

Secondary outcomes  

Blood loss and allogenic transfusion requirements
The blood loss was significantly lower in the thromboelastography group (MD -1.83 litre; 95% CI -2.55 to -1.12) and low CVP group (MD -1.12 litre; 95% CI -1.19 to -0.49). The network meta-analysis did not show any intervention to be effective in decreasing the proportion of patients requiring allogeneic red cell or whole blood transfusion. Convergence was not obtained for the allogeneic red cell or whole blood transfusion. The platelet transfusion was significantly lower in the tranexamic acid group (SMD -0.33; 95% CI -0.58 to -0.08). The network meta-analysis did not show any intervention to be effective in decreasing the amount of fresh frozen plasma transfused, or cryoprecipitate transfused.

Hospital stay
The network meta-analysis did not show any significant differences in the mean total hospital stay or ITU stay in the different interventions.

Variations in statistical analysis

Calculating the risk difference did not alter the interpretation of effectiveness of the interventions in any way for the binary outcomes.

Subgroup analysis

All trials were of high risk of bias. There were few trials included under each outcome and the details necessary for subgroup analysis was not available for many trials. So, subgroup analysis was not performed.

Sensitivity analysis

There was no alteration in the interpretation of effectiveness of the interventions by performing the various sensitivity analyses.

Assessment of reporting bias

Funnel plot was not performed because of the few trials that were included under each comparison for primary outcomes.

Trial sequential analysis

Trial sequential analysis was performed when there were at least two trials under the comparison. The comparisons for which trial sequential analysis was performed were aprotinin versus control, tranexamic acid versus control, recombinant factor VIIa versus control, and tranexamic acid versus aprotinin. The required information size was calculated using an alpha error of 0.05, power of 80%, control event proportion considered to be 10% based on an approximation of Eurpeon Liver Transplant Registry data (ELTR 2011), and a relative risk reduction of 20% for mortality and was adjusted for heterogeneity based on model variance using TSA 2011. The current information size, required information size, and the information fraction for the comparisons were 280/15,302 (1.83%) for aprotinin versus control, 139/18706 (0.74%) tranexamic acid versus control, 286/15.302 (1.87%) recombinant factor VIIa versus control, and 178/15,302 (1.16%) tranexamic acid versus aprotinin. This information was too little to represent the available information and draw a boundary for any of the comparisons (Figure 4; Figure 5; Figure 6; Figure 7).

Figure 4.

Trial sequential analysis (60-day mortality) - aprotinin versus control.

The blue line indicates the cumulative Z-curve.

Current information size: 280 patients; required information size: 15,302. information fraction: 1.83%.

Figure 5.

Trial sequential analysis (60-day mortality) - tranexamic acid versus control.

The blue line indicates the cumulative Z-curve.

Current information size: 139 patients; required information size: 18,706. information fraction: 0.74%.

Figure 6.

Trial sequential analysis (60-day mortality) - recombinant factor VIIa versus control.

The blue line indicates the cumulative Z-curve.

Current information size: 286 patients; required information size: 15,302. information fraction: 1.87%.

Figure 7.

Trial sequential analysis (60-day mortality) - tranexamic acid versus aprotinin.

The blue line indicates the cumulative Z-curve.

Current information size: 178 patients; required information size: 15,302. information fraction: 1.16%.

Discussion

Summary of main results

This systematic review was set out to evaluate all the interventions aimed at decreasing blood loss and transfusion requirements for liver transplantation. The direct meta-analysis did not identify any significant differences in the mortality in any comparisons. The network meta-analysis revealed a lower mortality in aprotinin and tranexamic acid groups. Thus, the results of direct analysis and network meta-analysis are potentially different. Such differences may be due to heterogeneity or bias (Salanti 2008). As discussed previously in this review, the trials were all at high risk of bias. In addition, the trials spanned over several years during which the surgical experience and the results of liver transplantation have improved (ELTR 2011), resulting in heterogeneity in the trials. Thus, these results are not as reliable as results from direct analysis, but warrant further investigation (Mills 2011).

Mortality at maximal follow-up was not reported in most trials. This was quite understandable considering that excessive blood loss affects mainly the peri-operative outcomes. However, recent evidence from observational studies suggest that increased peri-operative blood transfusion decreases long-term survival after liver transplantation (Ramos 2003; Massicotte 2005; Boyd 2007; Boin 2008). So, new trials assessing interventions aimed at decreasing the blood transfusion requirements after liver transplantation should also consider evaluating the long-term survival after liver transplantation.

There was no significant difference in the primary graft non-function, graft failure, or retransplantation in any of the comparisons. There were no significant differences in the thromboembolic episodes or in serious adverse events in any of the comparisons. However, the confidence intervals were wide, and it is not possible to reach any conclusion on the safety of the interventions.

None of the trials reported the quality of life in patients. This is one of the main aspects of healthcare decisions using economic modelling in state-funded healthcare systems. Future trials should include quality of life as an important outcome.

Overall completeness and applicability of evidence

None of the trials evaluated cell salvage method of decreasing allogeneic transfusion. Cell salvage may be an effective method of decreasing blood loss and transfusion requirements (Murphy 2004), it can be used in combination with anti-fibrinolytics (Diprose 2005), and should be evaluated in future trials in liver transplantation.

Most of the trials were performed more than a decade ago. The results of liver transplantation has steadily improved during the last 10 to 15 years (ELTR 2011), although there is no evidence from this review that the blood transfusion requirements have decreased over time. Aprotinin has been withdrawn from market because of concerns about increased risk of death in high risk cardiac procedures (MHRA 2007; Fergusson 2008; Ray 2008). The significantly higher deaths in the aprotinin group compared with tranexamic acid or aminocaproic acid in this trial was because of cardiac reasons with no significant difference in the deaths due to other causes. Thus, the increased risk of death might be specific to the population. We noted no such increase in the death, thromboembolic complications, or other serious adverse events in the arms involving aprotinin. So, evidence from well- designed randomised clinical trials (particularly those evaluating tranexamic acid, amino caproic acid, cell salvage, thromboelastography, and low central venous pressure) are necessary before any of the interventions can be advocated.

Quality of the evidence

The quality of evidence was generally poor as shown in the 'Summary of findings' tables (Summary of findings for the main comparison; Summary of findings 2; Summary of findings 3; Summary of findings 4).

There was inconsistency of results between direct meta-analysis and network meta-analysis probably due to bias and heterogeneity. The trial sequential analysis revealed that the number of patients included in the meta-analyses for 60-day mortality was very small compared with the optimal number of patients required to make any firm conclusions.

Potential biases in the review process

We imputed standard deviation when they were not available from trial reports or author replies. We imputed this as the maximum standard deviation from similar trials. The effect of such imputation is to decrease the weight of the study for calculation of mean differences and bias the effect estimate to no effect in case of standardised mean difference. We performed sensitivity analysis of using the average standard deviation, which did not alter the results. The alternative is to exclude these trials for these outcomes, which will make interpretation of the results from these trials even more difficult. We imputed the missing mean values from the median values. This might have introduced bias.

Agreements and disagreements with other studies or reviews

Previous reviews mainly assessed the role of antifibrinolytics and concluded that antifibrinolytics (mainly aprotinin) were safe and decreased blood loss although there was no reduction in the overall mortality or morbidity (Molenaar 2007; Liu 2008). By following a more rigorous assessment of the evidence, we conclude that aprotinin, tranexamic acid, recombinant factor VIIa, low central venous pressure, and thromboelastography may lower blood loss and transfusion requirements. A Cochrane review which assessed the role of antifibrinolytics in all surgeries concluded that there were no serious adverse events related to tranexamic acid and epsilon amino caproic acid and that aprotinin decreased the blood loss and transfusion requirements compared to the lysine analogues (tranexamic acid and epsilon amino caproic acid) but increased the risk of death (Henry 2011). However, the data related to death were obtained from the trial in which aprotinin increased risk of death due to cardiac causes in high-risk cardiac procedures (Fergusson 2008).

Authors' conclusions

Implications for practice

Aprotinin, recombinant factor VIIa, and thromboelastography groups may potentially lower blood loss and transfusion requirements. However, their benefit-harms ratios are not sufficiently studied due to the risk of systematic errors (bias) and the risk of random errors (play of chance).

Implications for research

Further well-designed trials with low risk of systematic error and low risk of random errors are necessary. Such trials should report the findings according to the CONSORT Statement (http://www.consort-statement.org/).

Acknowledgements

To The Cochrane Hepato-Biliary Group.

Peer Reviewers: Luit Penninga, Denmark; Christian Gluud, Denmark.
Contact Editor: Frederik Keus, The Netherlands.

Data and analyses

Download statistical data

Comparison 1. Intervention versus control
Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size
1 60-day mortality14 Risk Ratio (M-H, Fixed, 95% CI)Subtotals only
1.1 Aprotinin versus control3280Risk Ratio (M-H, Fixed, 95% CI)0.52 [0.18, 1.45]
1.2 Tranexamic acid versus control3139Risk Ratio (M-H, Fixed, 95% CI)0.55 [0.17, 1.76]
1.3 Epsilon amino caproic acid (EACA) versus control162Risk Ratio (M-H, Fixed, 95% CI)0.71 [0.13, 3.94]
1.4 Antithrombin III versus control129Risk Ratio (M-H, Fixed, 95% CI)0.24 [0.01, 4.65]
1.5 Recombinant factor VIIa (rFVIIa) versus control3286Risk Ratio (M-H, Fixed, 95% CI)1.51 [0.33, 6.95]
1.6 Aprotinin: bolus versus continuous infusion123Risk Ratio (M-H, Fixed, 95% CI)0.0 [0.0, 0.0]
1.7 Solvent detergent plasma versus standard fresh frozen plasma125Risk Ratio (M-H, Fixed, 95% CI)0.0 [0.0, 0.0]
1.8 Tranexamic acid versus EACA184Risk Ratio (M-H, Fixed, 95% CI)1.0 [0.21, 4.67]
1.9 Tranexamic acid versus aprotinin2178Risk Ratio (M-H, Fixed, 95% CI)4.12 [0.71, 23.76]
2 Mortality at maximal follow-up2 Risk Ratio (M-H, Fixed, 95% CI)Subtotals only
2.1 Antithrombin III versus control129Risk Ratio (M-H, Fixed, 95% CI)0.25 [0.03, 1.85]
2.2 Thromboelastography versus control128Risk Ratio (M-H, Fixed, 95% CI)0.67 [0.13, 3.40]
3 Primary graft non-function4 Risk Ratio (M-H, Fixed, 95% CI)Subtotals only
3.1 Aprotinin versus control2217Risk Ratio (M-H, Fixed, 95% CI)0.15 [0.02, 1.25]
3.2 Tranexamic acid versus control132Risk Ratio (M-H, Fixed, 95% CI)3.0 [0.13, 68.57]
3.3 Recombinant factor VIIa (rFVIIa) versus control120Risk Ratio (M-H, Fixed, 95% CI)0.0 [0.0, 0.0]
4 Retransplantation7 Risk Ratio (M-H, Random, 95% CI)Subtotals only
4.1 Aprotinin versus control2217Risk Ratio (M-H, Random, 95% CI)0.21 [0.02, 1.79]
4.2 Tranexamic acid versus control277Risk Ratio (M-H, Random, 95% CI)0.79 [0.18, 3.48]
4.3 Prostaglandin versus control120Risk Ratio (M-H, Random, 95% CI)0.0 [0.0, 0.0]
4.4 Aprotinin: bolus versus continuous infusion123Risk Ratio (M-H, Random, 95% CI)0.77 [0.05, 10.85]
4.5 Tranexamic acid versus aprotinin1127Risk Ratio (M-H, Random, 95% CI)2.95 [0.12, 71.17]
5 Thromboembolic episodes12 Risk Ratio (M-H, Fixed, 95% CI)Subtotals only
5.1 Aprotinin versus control3280Risk Ratio (M-H, Fixed, 95% CI)0.60 [0.18, 1.96]
5.2 Tranexamic acid versus control5179Risk Ratio (M-H, Fixed, 95% CI)2.20 [0.38, 12.64]
5.3 Epsilon amino caproic acid (EACA) versus control162Risk Ratio (M-H, Fixed, 95% CI)0.95 [0.09, 9.89]
5.4 Recombinant factor VIIa (rFVIIa) versus control2266Risk Ratio (M-H, Fixed, 95% CI)1.38 [0.65, 2.91]
5.5 Tranexamic acid versus EACA184Risk Ratio (M-H, Fixed, 95% CI)2.0 [0.39, 10.34]
5.6 Tranexamic acid versus aprotinin2178Risk Ratio (M-H, Fixed, 95% CI)1.97 [0.37, 10.37]
6 Serious adverse events4 Risk Ratio (M-H, Fixed, 95% CI)Subtotals only
6.1 Aprotinin versus control1137Risk Ratio (M-H, Fixed, 95% CI)1.02 [0.49, 2.11]
6.2 Recombinant factor VIIa (rFVIIa) versus control2266Risk Ratio (M-H, Fixed, 95% CI)1.30 [0.94, 1.78]
6.3 Low central venous pressure versus control186Risk Ratio (M-H, Fixed, 95% CI)0.71 [0.55, 0.92]
7 Blood loss11 Mean Difference (IV, Random, 95% CI)Subtotals only
7.1 Aprotinin versus control3195Mean Difference (IV, Random, 95% CI)-1.36 [-3.39, 0.66]
7.2 Tranexamic acid versus control265Mean Difference (IV, Random, 95% CI)-4.98 [-10.18, 0.23]
7.3 Antithrombin III versus control129Mean Difference (IV, Random, 95% CI)1.9 [-4.86, 8.66]
7.4 Thromboelastography versus control262Mean Difference (IV, Random, 95% CI)-1.13 [-1.85, -0.41]
7.5 Low central venous pressure versus control186Mean Difference (IV, Random, 95% CI)-1.19 [-1.88, -0.50]
7.6 Tranexamic acid versus aprotinin271Mean Difference (IV, Random, 95% CI)-1.01 [-2.31, 0.29]
8 Red-cell or whole blood transfusion4 Risk Ratio (M-H, Fixed, 95% CI)Subtotals only
8.1 Aprotinin versus control180Risk Ratio (M-H, Fixed, 95% CI)0.68 [0.40, 1.18]
8.2 Tranexamic acid versus control162Risk Ratio (M-H, Fixed, 95% CI)0.77 [0.60, 0.98]
8.3 Epsilon amino caproic acid (EACA) versus control162Risk Ratio (M-H, Fixed, 95% CI)0.90 [0.77, 1.06]
8.4 Recombinant factor VIIa (rFVIIa) versus control1182Risk Ratio (M-H, Fixed, 95% CI)0.91 [0.86, 0.97]
8.5 Tranexamic acid versus EACA184Risk Ratio (M-H, Fixed, 95% CI)0.81 [0.64, 1.02]
8.6 Tranexamic acid versus aprotinin1127Risk Ratio (M-H, Fixed, 95% CI)0.94 [0.71, 1.23]
9 Red cell or whole blood transfusion29 Std. Mean Difference (IV, Fixed, 95% CI)Subtotals only
9.1 Aprotinin versus control8375Std. Mean Difference (IV, Fixed, 95% CI)-0.61 [-0.82, -0.40]
9.2 Tranexamic acid versus control4159Std. Mean Difference (IV, Fixed, 95% CI)-0.27 [-0.59, 0.06]
9.3 Epsilon amino caproic acid (EACA) versus control162Std. Mean Difference (IV, Fixed, 95% CI)-0.17 [-0.71, 0.36]
9.4 Antithrombin III versus control129Std. Mean Difference (IV, Fixed, 95% CI)-0.04 [-0.77, 0.69]
9.5 Recombinant factor VIIa (rFVIIa) versus control2221Std. Mean Difference (IV, Fixed, 95% CI)-0.05 [-0.32, 0.23]
9.6 Oestrogen versus control130Std. Mean Difference (IV, Fixed, 95% CI)-0.58 [-1.32, 0.15]
9.7 Prostaglandin versus control120Std. Mean Difference (IV, Fixed, 95% CI)-0.04 [-0.92, 0.83]
9.8 Norepinephrine versus control165Std. Mean Difference (IV, Fixed, 95% CI)-0.04 [-0.53, 0.44]
9.9 Thromboelastography versus control262Std. Mean Difference (IV, Fixed, 95% CI)-0.73 [-1.25, -0.20]
9.10 Low central venous pressure versus control186Std. Mean Difference (IV, Fixed, 95% CI)-1.17 [-1.62, -0.71]
9.11 Aprotinin: bolus versus continuous infusion123Std. Mean Difference (IV, Fixed, 95% CI)0.86 [-0.00, 1.73]
9.12 Aprotinin: high dose versus medium dose1189Std. Mean Difference (IV, Fixed, 95% CI)0.16 [-0.12, 0.45]
9.13 Aprotinin: high dose versus low dose120Std. Mean Difference (IV, Fixed, 95% CI)-1.46 [-2.47, -0.45]
9.14 Whole blood versus blood components133Std. Mean Difference (IV, Fixed, 95% CI)-0.04 [-0.72, 0.65]
9.15 Solvent detergent plasma versus standard fresh frozen plasma125Std. Mean Difference (IV, Fixed, 95% CI)0.19 [-0.60, 0.98]
9.16 Tranexamic acid versus EACA184Std. Mean Difference (IV, Fixed, 95% CI)-0.23 [-0.66, 0.20]
9.17 Tranexamic acid versus aprotinin3198Std. Mean Difference (IV, Fixed, 95% CI)-0.09 [-0.36, 0.19]
10 Platelet transfusion21 Std. Mean Difference (IV, Fixed, 95% CI)Subtotals only
10.1 Aprotinin versus control5185Std. Mean Difference (IV, Fixed, 95% CI)-0.44 [-0.73, -0.14]
10.2 Tranexamic acid versus control4159Std. Mean Difference (IV, Fixed, 95% CI)-0.17 [-0.49, 0.16]
10.3 Epsilon amino caproic acid (EACA) versus control162Std. Mean Difference (IV, Fixed, 95% CI)0.01 [-0.52, 0.55]
10.4 Antithrombin III versus control129Std. Mean Difference (IV, Fixed, 95% CI)0.0 [-0.73, 0.73]
10.5 Recombinant factor VIIa (rFVIIa) versus control2216Std. Mean Difference (IV, Fixed, 95% CI)-0.03 [-0.31, 0.25]
10.6 Oestrogen versus control130Std. Mean Difference (IV, Fixed, 95% CI)-0.53 [-1.26, 0.20]
10.7 Norepinephrine versus control165Std. Mean Difference (IV, Fixed, 95% CI)0.34 [-0.15, 0.83]
10.8 Thromboelastography versus control128Std. Mean Difference (IV, Fixed, 95% CI)-0.17 [-0.91, 0.58]
10.9 Aprotinin: high dose versus medium dose1189Std. Mean Difference (IV, Fixed, 95% CI)-0.48 [-0.77, -0.19]
10.10 Whole blood versus blood components133Std. Mean Difference (IV, Fixed, 95% CI)-0.12 [-0.81, 0.56]
10.11 Solvent detergent plasma versus standard fresh frozen plasma125Std. Mean Difference (IV, Fixed, 95% CI)0.07 [-0.71, 0.86]
10.12 Tranexamic acid versus EACA184Std. Mean Difference (IV, Fixed, 95% CI)0.05 [-0.37, 0.48]
10.13 Tranexamic acid versus aprotinin3198Std. Mean Difference (IV, Fixed, 95% CI)-0.08 [-0.36, 0.20]
11 Plasma transfusion25 Std. Mean Difference (IV, Fixed, 95% CI)Subtotals only
11.1 Aprotinin versus control8420Std. Mean Difference (IV, Fixed, 95% CI)-0.33 [-0.53, -0.13]
11.2 Tranexamic acid versus control4159Std. Mean Difference (IV, Fixed, 95% CI)-0.21 [-0.53, 0.11]
11.3 Epsilon amino caproic acid (EACA) versus control162Std. Mean Difference (IV, Fixed, 95% CI)-0.12 [-0.66, 0.41]
11.4 Antithrombin III versus control129Std. Mean Difference (IV, Fixed, 95% CI)0.04 [-0.69, 0.77]
11.5 Recombinant factor VIIa (rFVIIa) versus control2262Std. Mean Difference (IV, Fixed, 95% CI)-0.01 [-0.28, 0.25]
11.6 Oestrogen versus control130Std. Mean Difference (IV, Fixed, 95% CI)-0.67 [-1.41, 0.07]
11.7 Prostaglandin versus control120Std. Mean Difference (IV, Fixed, 95% CI)0.0 [-0.88, 0.88]
11.8 Norepinephrine versus control165Std. Mean Difference (IV, Fixed, 95% CI)0.11 [-0.38, 0.60]
11.9 Thromboelastography versus control128Std. Mean Difference (IV, Fixed, 95% CI)-0.82 [-1.60, -0.05]
11.10 Aprotinin: bolus versus continuous infusion123Std. Mean Difference (IV, Fixed, 95% CI)0.05 [-0.78, 0.87]
11.11 Aprotinin: high dose versus medium dose1189Std. Mean Difference (IV, Fixed, 95% CI)0.06 [-0.22, 0.35]
11.12 Aprotinin: high dose versus low dose120Std. Mean Difference (IV, Fixed, 95% CI)-1.31 [-2.29, -0.32]
11.13 Whole blood versus blood components133Std. Mean Difference (IV, Fixed, 95% CI)-1.53 [-2.32, -0.74]
11.14 Tranexamic acid versus EACA184Std. Mean Difference (IV, Fixed, 95% CI)-0.22 [-0.65, 0.21]
11.15 Tranexamic acid versus aprotinin271Std. Mean Difference (IV, Fixed, 95% CI)0.02 [-0.45, 0.48]
12 Cryoprecipitate12 Std. Mean Difference (IV, Fixed, 95% CI)Subtotals only
12.1 Aprotinin versus control3147Std. Mean Difference (IV, Fixed, 95% CI)-0.49 [-0.82, -0.16]
12.2 Tranexamic acid versus control4181Std. Mean Difference (IV, Fixed, 95% CI)-0.27 [-0.57, 0.02]
12.3 Epsilon amino caproic acid (EACA) versus control162Std. Mean Difference (IV, Fixed, 95% CI)0.00 [-0.53, 0.54]
12.4 Antithrombin III versus control129Std. Mean Difference (IV, Fixed, 95% CI)0.0 [-0.73, 0.73]
12.5 Oestrogen versus control130Std. Mean Difference (IV, Fixed, 95% CI)0.0 [-0.72, 0.72]
12.6 Thromboelastography versus control128Std. Mean Difference (IV, Fixed, 95% CI)-0.25 [1.00, 0.49]
12.7 Whole blood versus blood components133Std. Mean Difference (IV, Fixed, 95% CI)-0.28 [-0.97, 0.41]
12.8 Solvent detergent plasma versus standard fresh frozen plasma125Std. Mean Difference (IV, Fixed, 95% CI)0.10 [-0.68, 0.89]
12.9 Tranexamic acid versus EACA162Std. Mean Difference (IV, Fixed, 95% CI)-0.15 [-0.69, 0.38]
13 Hospital stay6 Mean Difference (IV, Fixed, 95% CI)Subtotals only
13.1 Aprotinin versus control163Mean Difference (IV, Fixed, 95% CI)0.0 [-4.94, 4.94]
13.2 Tranexamic acid versus control145Mean Difference (IV, Fixed, 95% CI)4.0 [-16.18, 24.18]
13.3 Recombinant factor VIIa (rFVIIa) versus control1180Mean Difference (IV, Fixed, 95% CI)3.40 [-7.51, 14.31]
13.4 Norepinephrine versus control165Mean Difference (IV, Fixed, 95% CI)-1.00 [-3.18, 1.18]
13.5 Low central venous pressure versus control186Mean Difference (IV, Fixed, 95% CI)0.0 [-19.43, 19.43]
13.6 Tranexamic acid versus aprotinin151Mean Difference (IV, Fixed, 95% CI)5.0 [-2.69, 12.69]
14 Intensive therapy unit stay9 Mean Difference (IV, Fixed, 95% CI)Subtotals only
14.1 Aprotinin versus control2199Mean Difference (IV, Fixed, 95% CI)-0.21 [-0.78, 0.36]
14.2 Tranexamic acid versus control145Mean Difference (IV, Fixed, 95% CI)-1.0 [-4.75, 2.75]
14.3 Recombinant factor VIIa (rFVIIa) versus control2199Mean Difference (IV, Fixed, 95% CI)0.22 [-0.14, 0.58]
14.4 Norepinephrine versus control165Mean Difference (IV, Fixed, 95% CI)-0.13 [-1.65, 1.39]
14.5 Low central venous pressure versus control186Mean Difference (IV, Fixed, 95% CI)-1.0 [-3.71, 1.71]
14.6 Aprotinin: bolus versus continuous infusion123Mean Difference (IV, Fixed, 95% CI)2.0 [-3.99, 7.99]
14.7 Tranexamic acid versus aprotinin151Mean Difference (IV, Fixed, 95% CI)0.0 [-2.40, 2.40]
Analysis 1.1.

Comparison 1 Intervention versus control, Outcome 1 60-day mortality.

Analysis 1.2.

Comparison 1 Intervention versus control, Outcome 2 Mortality at maximal follow-up.

Analysis 1.3.

Comparison 1 Intervention versus control, Outcome 3 Primary graft non-function.

Analysis 1.4.

Comparison 1 Intervention versus control, Outcome 4 Retransplantation.

Analysis 1.5.

Comparison 1 Intervention versus control, Outcome 5 Thromboembolic episodes.

Analysis 1.6.

Comparison 1 Intervention versus control, Outcome 6 Serious adverse events.

Analysis 1.7.

Comparison 1 Intervention versus control, Outcome 7 Blood loss.

Analysis 1.8.

Comparison 1 Intervention versus control, Outcome 8 Red-cell or whole blood transfusion.

Analysis 1.9.

Comparison 1 Intervention versus control, Outcome 9 Red cell or whole blood transfusion.

Analysis 1.10.

Comparison 1 Intervention versus control, Outcome 10 Platelet transfusion.

Analysis 1.11.

Comparison 1 Intervention versus control, Outcome 11 Plasma transfusion.

Analysis 1.12.

Comparison 1 Intervention versus control, Outcome 12 Cryoprecipitate.

Analysis 1.13.

Comparison 1 Intervention versus control, Outcome 13 Hospital stay.

Analysis 1.14.

Comparison 1 Intervention versus control, Outcome 14 Intensive therapy unit stay.

Appendices

Appendix 1. Search strategies for identification of studies

DatabasePeriod of SearchSearch Strategy
Cochrane Central Register of Controlled Trials (CENTRAL) in The Cochrane Library (Wiley)Issue 3, 2011.#1 Blood loss OR bleeding OR hemorrhage OR haemorrhage OR hemorrhages OR haemorrhages OR hemostasis OR haemostasis OR transfusion
#2 MeSH descriptor Hemorrhage explode all trees
#3 MeSH descriptor Blood Transfusion explode all trees
#4 (#1 OR #2 OR #3)
#5 liver OR hepatic OR hepato
#6 MeSH descriptor Liver explode all trees
#7 (#5 OR #6)
#8 resection OR segmentectomy OR graft* OR transplant*
#9 (#7 AND #8)
#10 hepatectomy
#11 MeSH descriptor Hepatectomy explode all trees
#12 MeSH descriptor Liver Transplantation explode all trees
#13 (#9 OR #10 OR #11 OR #12)
#14 (#4 AND #13)
MEDLINE (Pubmed)1990 to September 2011.(Blood loss OR bleeding OR hemorrhage OR haemorrhage OR hemorrhages OR haemorrhages OR hemostasis OR haemostasis OR transfusion OR "Hemorrhage"[Mesh] OR "Blood Transfusion"[Mesh]) AND (((liver OR hepatic OR hepato OR “liver”[MeSH]) AND (resection OR segmentectomy OR graft* OR transplant*)) OR hepatectomy OR “hepatectomy”[MeSH] OR “Liver Transplantation”[MeSH]) AND (((randomised controlled trial [pt] OR controlled clinical trial [pt] OR randomised controlled trials [mh] OR random allocation [mh] OR double-blind method [mh] OR single-blind method [mh] OR clinical trial [pt] OR clinical trials [mh] OR ("clinical trial" [tw]) OR ((singl* [tw] OR doubl* [tw] OR trebl* [tw] OR tripl* [tw]) AND (mask* [tw] OR blind* [tw])) OR (placebos [mh] OR placebo* [tw] OR random* [tw] OR research design [mh:noexp]) NOT (animals [mh] NOT human [mh]))))
EMBASE (OvidSP)1990 to September 2011.1 (blood loss or bleeding or hemorrhage or haemorrhage or hemorrhages or haemorrhages or hemostasis or haemostasis or transfusion).af. 
2 exp bleeding/ or exp blood transfusion/ 
3 1 or 2 
4 (liver or hepatic or hepato).af. 
5 (segmentectomy or resection or graft* or transplant*).af. 
6 4 and 5 
7 hepatectomy.af. 
8 exp liver resection/ or exp liver transplantation/ 
9 6 or 7 or 8 
10 3 and 9 
11 exp crossover-procedure/ or exp double-blind procedure/ or exp randomised controlled trial/ or single-blind procedure/ 
12 (random* or factorial* or crossover* or placebo*).af. 
13 11 or 12 
14 10 and 13
Science Citation Index Expanded (http://portal.isiknowledge.com/portal.cgi?DestApp=WOS&Func=Frame)1990 to September 2011.#1 TS=(Blood loss OR bleeding OR hemorrhage OR haemorrhage OR hemorrhages OR haemorrhages OR hemostasis OR haemostasis OR transfusion)
#2 TS=(((liver OR hepatic OR hepato) AND (resection OR segmentectomy OR graft* OR transplant*)) OR hepatectomy)
#3 TS=(random* OR blind* OR placebo* OR meta-analysis)
#4 #3 AND #2 AND #1
metaRegister of Controlled Trials (Current Controlled Trials)September 2011.(transfusion or bleeding or hemorrhage or haemorrhage) AND (liver transplant* or liver graft*)

Appendix 2. Sample SAS code for network meta-analysis

Sample codeType of outcome

data NetworkMetaAnalysis;

input trial   x              n           arm1      arm3      arm6      arm7       arm9;

datalines;                                                                                                           

3              0              25           0              0              1              0              0

3              0              20           1              0              0              0              0

5              4              42           0              0              1              0              0

5              2              42           0              0              0              1              0

5              2              40           1              0              0              0              0

6              4              64           0              0              1              0              0

6              2              63           0              1              0              0              0

8              1              33           0              1              0              0              0

8              2              30           1              0              0              0              0

10            1              39           0              1              0              0              0

10            0              41           1              0              0              0              0

17            0              24           0              1              0              0              0

17            0              27           0              0              1              0              0

18            1              16           0              0              1              0              0

18            0              16           1              0              0              0              0

21            7              58           0              0              0              0              1

21          12             63           0              0              0              0              1

21            6              62           1              0              0              0              0

25            2              18           0              0              0              0              1

25            1              24           0              0              0              0              1

25            2              22           0              0              0              0              1

25            2              19           1              0              0              0              0

27            2              46           0              1              0              0              0

27            0              43           0              1              0              0              0

27            3              48           1              0              0              0              0

28            0              10           0              0              0              0              1

28            0              10           1              0              0              0              0

33            0              10           0              0              1              0              0

33            0              10           1              0              0              0              0

run;
* the parameters for random effects model can be calculated from simple logistic regression;
* code for simple logistic regression;
proc logistic data=NetworkMetaAnalysis;
class arm;
model x/n = arm;
run;
* random effects for trial and treatment contrasts;
proc nlmixed data=NetworkMetaAnalysis;
parms base=-1 s2trialr=1 s2het=1 i=-2.9109 a=-0.0194 c=-0.7863 f=-0.0567 g=-0.0849;
logitp = (base + trialr) + arm9*(i+het) + arm3*(c+het) + arm6*(f+het) + arm7*(g+het);
P = exp(logitp)/(1+exp(logitp));
model x ˜ binomial(n,P);
random trialr het ˜ normal([0,0],[s2trialr,0,s2het])
subject=trial;
run

Binary outcome

data NetworkMetaAnalysis;
/* paste header row after input*/

input trial  mean  n  MaxSD  AvgSD   arm1      arm3      arm6      arm9      arm18   arm19;

/*paste the data after the datalines; */

datalines;

3           28  25   27        27          0             0             1           0               0       0

3           24   20    39.2      39.2        1             0             0           0               0       0

7           35   43     23         23        0             0             0           0               1       0

7           35   43    60.8      60.8      1              0             0           0              0       0

8             9   33      6.8        6.8        0             1            0           0               0       0

8             9   30    12.2      12.2        1             0            0           0      0    0

17           26  24      7            7        0              1             0           0               0       0

17           31  27    19         19         0              0            1           0               0       0

21           22   56    38.2       18.4         0             0            0           1               0       0

21           19   62       33      15.9         0              0            0           1               0       0

21           17   61    29.5       14.2          1              0            0           0               0      0

26           7.8   33   4.96      4.96        0              0            0           0               0       1

26           8.8   32   3.98      3.98         1             0            0           0              0       0

run;
title1 'Hospital stay - Maximum standard deviation';
proc nlmixed data=NetworkMetaAnalysis;
parms base=-1 s2trialr=1 s2het=1;
y = (base + trialr) + arm3*(c+het)+ arm6*(f+het)+ arm9*(i+het)+ arm18*(r+het)+ arm19*(s+het);
se2 = MaxSD**2/n;
model mean ˜ normal(y,se2);
random trialr het ˜ normal([0,0],[s2trialr,0,s2het])
subject=trial;
run

Continuous outcome

What's new

DateEventDescription
20 March 2012AmendedSpelling mistake in the name of Luit Penninga, peer reviewer on the review, is corrected.

Contributions of authors

KS Gurusamy wrote the review, assessed the trials for inclusion, extracted data, performed data analysis, and wrote the draft of the review. T Pissanou assessed the trials for inclusion and extracted data independently. H Pikhart helped with the analysis of data. J Vaughan improved the readability of the review and clarified any unclear issues. AK Burroughs and BR Davidson critically commented on the review. All the authors approve of the present review.

Declarations of interest

None.

Sources of support

Internal sources

  • Department of Surgery, University College London, UK.

External sources

  • NIHR Cochrane Incentive Scheme, UK.

    A grant of £5000 by the UK Government health research funding body (National Insititute of Health Research or NIHR) to perform this review.

Differences between protocol and review

  1. The risk of bias domains were updated according to The Cochrane Handbook for Systematic Reviews of Interventions Version 5.1.0 (Higgins 2011).

  2. Network analysis was performed in addition to routine meta-analysis.

  3. Fisher's exact test was used when there was only one trial included in the comparison of a binary outcome between two groups.

  4. In the protocol we mentioned that the proportion of patients who developed mortality in the control group will be based on the overall proportion of patients who died in this systematic review. However the trials included included those more than 15 years old and there has been significant decrease in the mortality over this period (ELTR 2011). So, we based our trial sequential calculations based on an approximation of the ELTR data (ELTR 2011).

Characteristics of studies

Characteristics of included studies [ordered by study ID]

Ahn 2008

MethodsRandomised clinical trial.
ParticipantsCountry: Korea.
Sample size: 74.
Post-randomisation drop-out(s): 8 (10.8%).
Revised sample size: 66.
Females: 12 (16.2%).
Mean age: 51 years.
Piggyback: not stated.
Conventional: not stated.
Cadaveric donor: not stated.
Live donor: not stated.
Inclusion criteria:
1. Adult patients undergoing primary orthotopic liver transplantation.
Exclusion criteria:
1. Severe cardiovascular, cerebrovascular, pulmonary, or end-stage renal diseases.
2. Taking medications likely to alter coagulation within two weeks of the study.
3. Allergic reaction to HES.
InterventionsThe patients were randomised to the following groups.
Group 1: intervention (n = 31).
Further details: high molecular weight hydroxy ethyl starch.
Group 2: control (n = 35).
Further details: low molecular weight hydroxy ethyl starch.
OutcomesThe outcomes reported were thromboelastography parameters. None of the outcomes included in this review were reported.
NotesAttempts were made to contact the author in September 2011.

Reason for post-randomisation drop-out(s): sampling error or transfusion with blood products.
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Unclear riskComment: This information was not available.
Allocation concealment (selection bias)Unclear riskComment: This information was not available.
Blinding of participants and personnel (performance bias)
All outcomes
Unclear riskComment: This information was not available.
Blinding of outcome assessment (detection bias)
All outcomes
Unclear riskComment: This information was not available.
Incomplete outcome data (attrition bias)
All outcomes
High riskComment: There were post-randomisation drop-outs.
Selective reporting (reporting bias)High riskComment: Important outcomes were not reported.
Vested interest biasUnclear riskComment: This information was not available.

Baudo 1992

MethodsRandomised clinical trial.
ParticipantsCountry: Italy.
Sample size: 29.
Post-randomisation drop-out(s): not stated.
Revised sample size: 29.
Females: 10 (34.5%).
Mean age: 42.9 years.
Piggyback: not stated.
Conventional: not stated.
Cadaveric donor: not stated.
Live donor: not stated.
Inclusion criteria:
Patients undergoing orthotopic liver transplantation.
InterventionsThe patients were randomised to the following groups.
Group 1: intervention (n = 13).
Further details: antithrombin III bolus on induction (100 - plasma activity level) x kg body weight followed by continuous infusion of 1000 units per hour.
Group 2: control (n = 16).
Further details: control (no intervention).
OutcomesThe outcomes reported were mortality, blood loss, and blood transfusion requirements.
NotesAttempts were made to contact the author in September 2011.
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Unclear riskComment: This information was not available.
Allocation concealment (selection bias)Unclear riskComment: This information was not available.
Blinding of participants and personnel (performance bias)
All outcomes
Unclear riskComment: This information was not available.
Blinding of outcome assessment (detection bias)
All outcomes
Unclear riskComment: This information was not available.
Incomplete outcome data (attrition bias)
All outcomes
Unclear riskComment: This information was not available.
Selective reporting (reporting bias)High riskComment: Important outcomes were not reported.
Vested interest biasUnclear riskComment: This information was not available.

Boylan 1996

MethodsRandomised clinical trial.
ParticipantsCountry: Canada.
Sample size: 45.
Post-randomisation drop-out(s): not stated.
Revised sample size: 45.
Females: not stated.
Mean age: 49.2 years.
Piggyback: not stated.
Conventional: not stated.
Cadaveric donor: not stated.
Live donor: not stated.
Inclusion criteria:
1. Primary isolated orthotopic liver transplantation between 1992 and 1994.
Exclusion criteria:
1. Patients with primary biliary cirrhosis or primary sclerosing cholangitis.
2. Predisposition to thrombotic tendency.
3. Fulminant liver failure.
InterventionsThe patients were randomised to the following groups.
Group 1: intervention (n = 25).
Further details: tranexamic acid 40 mg/kg/hour continuous infusion from induction of anaesthesia to unclamping of portal vein. maximum dose = 20 gram.
Group 2: control (n = 20).
Further details: normal saline.
OutcomesThe outcomes reported were mortality, retransplantation, hospital stay, and transfusion requirements.
NotesAttempts were made to contact the author in September 2011.
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Low riskQuote: "Study agents were prepared by the hospital pharmacy using a randomisation schedule provided in sealed envelopes".
Allocation concealment (selection bias)Low riskQuote: "Study agents were prepared by the hospital pharmacy using a randomisation schedule provided in sealed envelopes; all other personnel were blinded to randomisation status".
Comment: Probably adequate allocation concealment, although details of the sealed envelope such as whether they were opaque and consecutively numbered were not available.
Blinding of participants and personnel (performance bias)
All outcomes
Low riskQuote: "Study agents were prepared by the hospital pharmacy using a randomisation schedule provided in sealed envelopes; all other personnel were blinded to randomisation status".
Comment: Normal saline was used as placebo.
Blinding of outcome assessment (detection bias)
All outcomes
Low riskQuote: "Study agents were prepared by the hospital pharmacy using a randomisation schedule provided in sealed envelopes; all other personnel were blinded to randomisation status".
Comment: Normal saline was used as placebo.
Incomplete outcome data (attrition bias)
All outcomes
High riskComment: In a preliminary report, the authors had excluded two patients out of 30 patients.
Selective reporting (reporting bias)High riskComment: Important outcomes were not reported.
Vested interest biasUnclear riskComment: This information was not available.

Cottam 1991

MethodsRandomised clinical trial.
ParticipantsCountry: UK.
Sample size: 8.
Post-randomisation drop-out(s): not stated.
Revised sample size: 8.
Females: not stated.
Mean age: not stated.
Piggyback: not stated.
Conventional: not stated.
Cadaveric donor: not stated.
Live donor: not stated.
Inclusion criteria:
Patients undergoing liver transplantation
InterventionsThe patients were randomised to the following groups.
Group 1: intervention (n = 4).
Further details: aprotinin 2 million KIU at induction followed by 500,000 KIU per hour and 50,000 KIU for every unit of blood transfused.
Group 2: control (n = 4).
Further details: no intervention.
OutcomesThe outcomes reported were fibrinolysis parameters. None of the outcomes of interest for this review were included in the report.
NotesAttempts were made to contact the author in September 2011.
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Unclear riskComment: This information was not available.
Allocation concealment (selection bias)Unclear riskComment: This information was not available.
Blinding of participants and personnel (performance bias)
All outcomes
Unclear riskComment: This information was not available.
Blinding of outcome assessment (detection bias)
All outcomes
Unclear riskComment: This information was not available.
Incomplete outcome data (attrition bias)
All outcomes
Unclear riskComment: This information was not available.
Selective reporting (reporting bias)High riskComment: Important outcomes were not reported.
Vested interest biasUnclear riskComment: This information was not available.

Dalmau 2000

MethodsRandomised clinical trial.
ParticipantsCountry: Spain.
Sample size: 153.
Post-randomisation drop-out(s): 29 (19%).
Revised sample size: 124.
Females: 45 (29.4%).
Mean age: 58 years.
Piggyback: 124 (81%).
Conventional: 0 (0%).
Cadaveric donor: not stated.
Live donor: not stated.
Inclusion criteria:
1. Patients undergoing orthotopic liver transplantation.
Exclusion criteria:
1. Patients with Budd Chiari syndrome.
2. Acute hepatic failure.
3. Retransplantation of less than 1 month.
4. Simultaneous liver and kidney transplantation.
5. Amyloidosis.
InterventionsThe patients were randomised to the following groups.
Group 1: intervention1 (n = 42).
Further details of intervention: tranexamic acid at 10 mg/kg/hour from the anaesthesia induction to graft reperfusion.
Group 2: intervention2 (n = 42).
Further details of intervention: epsilon aminocaproic acid at 16 mg/kg/hour from the anaesthesia induction to graft reperfusion.
Group 3: control (n = 40).
Further details: normal saline as placebo.
OutcomesThe outcomes reported were mortality, thrombotic events, and transfusion requirements.
NotesAttempts were made to contact the author in September 2011.

Reason for post-randomisation drop-out(s): acute hepatic failure (3), retransplantation (11), simultaneous kidney and liver transplantations (3), renal insufficiency with dialysis (1), primary amyloidotic neuropathy (3), incomplete data (5), and air embolism with cardiac arrest (3).
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Unclear riskComment: This information was not available.
Allocation concealment (selection bias)Unclear riskComment: This information was not available.
Blinding of participants and personnel (performance bias)
All outcomes
Unclear riskComment: Although a placebo was used, the nature of the placebo was not stated.
Blinding of outcome assessment (detection bias)
All outcomes
Unclear riskComment: Although a placebo was used, the nature of the placebo was not stated.
Incomplete outcome data (attrition bias)
All outcomes
High riskComment: There were post-randomisation drop-outs.
Selective reporting (reporting bias)High riskComment: Important outcomes were not reported.
Vested interest biasUnclear riskComment: This information was not available.

Dalmau 2004

MethodsRandomised clinical trial.
ParticipantsCountry: Spain.
Sample size: 127.
Post-randomisation drop-out(s): 0 (0%).
Revised sample size: 127.
Females: 38 (29.9%).
Mean age: 53.5 years.
Piggyback: 127 (100%).
Conventional: not stated.
Cadaveric donor: not stated.
Live donor: not stated.
Inclusion criteria:
1. Patients undergoing orthotopic liver transplantation.
Exclusion criteria:
1. Patients with Budd Chiari syndrome.
2. Acute hepatic failure.
3. Retransplantation of less than 1 month.
4. Simultaneous liver and kidney transplantation.
5. Amyloidosis.
InterventionsThe patients were randomised to the following groups.
Group 1: intervention (n = 64).
Further details: tranexamic acid bolus at induction (quantity not stated) followed by continuous infusion of 10 mg/kg/hour until 2 hours after unclamping of portal vein.
Group 2: control (n = 63).
Further details: aprotinin bolus of 2 million ki units at induction followed by continuous infusion of 500,000 KIU/hour until 2 hours after portal vein unclamping.
OutcomesThe outcomes reported were mortality, retransplantation, reoperation, complications, and blood transfusion requirements.
NotesAttempts were made to contact the author in September 2011.
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Low riskQuote: "Drugs were prepared using a randomisation schedule provided in sealed envelopes".
Allocation concealment (selection bias)Low riskQuote: "Drugs were prepared using a randomisation schedule provided in sealed envelopes. The anesthesiologist, nurse, and surgeons were unaware of the details of the randomisation".
Comment: Further details of sealed envelopes were not available. However, based on the steps taken to conceal the drug from the health providers, the sealed envelopes were probably adequate. .
Blinding of participants and personnel (performance bias)
All outcomes
Low riskQuote: "Both drugs were diluted in normal saline in order to administer them at a rate of 100 ml/h after the bolus dose... The anesthesiologist, nurse, and surgeons were unaware of the details of the randomisation".
Blinding of outcome assessment (detection bias)
All outcomes
Low riskQuote: "Both drugs were diluted in normal saline in order to administer them at a rate of 100 ml/h after the bolus dose".. The anesthesiologist, nurse, and surgeons were unaware of the details of the randomisation".
Incomplete outcome data (attrition bias)
All outcomes
Low riskComment: There were no post-randomisation drop-outs.
Selective reporting (reporting bias)High riskComment: Important outcomes were not reported.
Vested interest biasLow riskQuote: "This study has not been the recipient of any grant or other financial support".

Feng 2010

MethodsRandomised clinical trial.
ParticipantsCountry: China.
Sample size: 86.
Post-randomisation drop-out(s): not stated.
Revised sample size: 86.
Females: 14 (16.3%).
Mean age: 47.5 years.
Piggyback: 86 (100%).
Conventional: 0 (0%).
Cadaveric donor: not stated.
Live donor: not stated.
Inclusion criteria:
1. Patients scheduled to undergo primary liver transplantation.
2. Age above 18 years.
Exclusion criteria:
1. Retransplantation.
2. Split liver transplantation.
3. Intraoperative demonstration of extended portal thrombosis.
InterventionsThe patients were randomised to the following groups.
Group 1: intervention (n = 43).
Further details: low central venous pressure (CVP kept below 5 mm hg or 40% below the baseline value) with mean arterial pressure at more than 60 mm Hg  by limiting infusion volume, adjusting posture with a 5 to 10 degree head-up tilt (reverse-Trendelenberg patient position), and infusion of somatostatin and nitroglycerin (0.5–3.0 mcg/min/kg).
Group 2: control (n = 43).
Further details: standard management.
OutcomesThe outcomes reported were complications, hospital stay, blood loss, and blood transfusion requirements.
NotesAuthor replied to questions related to bias in October 2011.
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Low risk

Quote: 86 opaque sealed envelopes including 43 envelopes with the number ¡°1¡± and another 43 envelopes with the number ¡°2¡± were put into one locked box. Once the patient was chosen for protocol and informed consent was obtained, a staff took out one envelope from the locked box randomly. After opening the envelope, the number ¡°1¡± meant the patient was divided into ¡° the LCVP group¡±, while the number ¡°2¡± meant the patient was divided into ¡°the control group¡±.

Comment: This method is similar to shuffling. This information was provided by author's replies to questions.

Allocation concealment (selection bias)Low risk

Quote: 86 opaque sealed envelopes including 43 envelopes with the number ¡°1¡± and another 43 envelopes with the number ¡°2¡± were put into one locked box. Once the patient was chosen for protocol and informed consent was obtained, a staff took out one envelope from the locked box randomly. After opening the envelope, the number ¡°1¡± meant the patient was divided into ¡° the LCVP group¡±, while the number ¡°2¡± meant the patient was divided into ¡°the control group¡±.

Comment: This method is similar to shuffling. This information was provided by author's replies to questions.

Blinding of participants and personnel (performance bias)
All outcomes
Unclear riskComment: This information was not available.
Blinding of outcome assessment (detection bias)
All outcomes
Unclear riskComment: This information was not available.
Incomplete outcome data (attrition bias)
All outcomes
Low risk

Quote: "There were no post-randomisation drop-outs".

Comment: This information was provided by author's replies to questions.

Selective reporting (reporting bias)High riskComment: Important outcomes were not reported.
Vested interest biasLow risk

Quote: "This study was supported by grants from the National Science and Technology (S & T) major project, China (2008ZX10002-026), and the Science and Technology Bureau of Zhejiang Province, China (2009C33091)."

Comment: This information was provided by author's replies to questions.

Findlay 2001

MethodsRandomised clinical trial.
ParticipantsCountry: USA.
Sample size: 63.
Post-randomisation drop-out(s): not stated.
Revised sample size: 63.
Females: 31 (49.2%).
Mean age: 51 years.
Piggyback: not stated.
Conventional: not stated.
Cadaveric donor: not stated.
Live donor: not stated.
Inclusion criteria:
1. Patients undergoing primary orthotopic liver transplantation.
Exclusion criteria:
1. Pre-OLT diagnosis of hepatorenal syndrome.
2. Established renal failure.
InterventionsThe patients were randomised to the following groups.
Group 1: intervention (n = 30).
Further details: aprotinin - loading dose 1 million KIU over 30 minutes; followed by an infusion of 250,000 KIU/hour until skin closure.
Group 2: control (n = 33).
Further details: placebo (normal saline).
OutcomesThe outcomes reported were mortality, serious adverse events, hospital stay, transfusion requirements.
NotesAuthor replied to questions related to risk of bias in October 2011.
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Low risk

Quote: "Randomization was by a stratified randomization scheme with 2 factors - surgeon (one for each of 3 surgeons) and diagnosis (hepatitis or non-hepatitis) for a total of 6 strata (3 surgeons x 2 diagnostic groups).  For each stratum a blocked randomization list was produced by the Biostatistics group and maintained by the pharmacy.  When a subject came to transplant pharmacy determined the stratum, selected the next envelope for that stratum and then issued the indicated trial medication (aprotinin or placebo) to the operating room".

Comment: This information was provided by author's replies to questions.

Allocation concealment (selection bias)Low risk

Quote: "Randomization was by a stratified randomization scheme with 2 factors - surgeon (one for each of 3 surgeons) and diagnosis (hepatitis or non-hepatitis) for a total of 6 strata (3 surgeons x 2 diagnostic groups).  For each stratum a blocked randomization list was produced by the Biostatistics group and maintained by the pharmacy.  When a subject came to transplant pharmacy determined the stratum, selected the next envelope for that stratum and then issued the indicated trial medication (aprotinin or placebo) to the operating room".

Comment: This information was provided by author's replies to questions.

Blinding of participants and personnel (performance bias)
All outcomes
Low riskQuote: "Subjects were randomised to the administration of either aprotinin or placebo (an equivalent infusion of normal saline)".
Blinding of outcome assessment (detection bias)
All outcomes
Low riskQuote: "Subjects were randomised to the administration of either aprotinin or placebo (an equivalent infusion of normal saline)".
Incomplete outcome data (attrition bias)
All outcomes
Low risk

Quote: There were no post-randomisation drop-outs.

Comment: This information was provided by author's replies to questions..

Selective reporting (reporting bias)High riskComment: Important outcomes were not reported.
Vested interest biasHigh risk

Quote: Funding was partially by a grant from Bayer Pharmaceuticals (as noted in the original publication), partially from internal Mayo departmental funding.

Comment: This information was provided by author's replies to questions..

Frenette 1998

MethodsRandomised clinical trial.
ParticipantsCountry: USA.
Sample size: 30.
Post-randomisation drop-out(s): not stated.
Revised sample size: 30.
Females: not stated.
Mean age: 49.5 years.
Piggyback: not stated.
Conventional: not stated.
Cadaveric donor: not stated.
Live donor: not stated.
Inclusion criteria:
1. Patients undergoing orthotopic liver transplantation.
2. Age more than 18 years.
3. Patients having a reaction time (R-time) more than 15 minutes.
Exclusion criteria:
1. Patients already taking oestrogen therapy.
2. Received blood products in preparation for surgery.
3. Patients undergoing retransplantation.
InterventionsThe patients were randomised to the following groups.
Group 1: intervention (n = 15).
Further details: conjugated oestrogen 100 mg intravenous when the reaction time was longer than 15 minutes at the beginning of surgery and 30 minutes after reperfusion of the new graft.
Group 2: control (n = 15).
Further details: placebo (equal volume of normal saline).
OutcomesThe outcomes reported were blood transfusion requirements.
NotesAttempts were made to contact the author in September 2011.
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Low riskQuote: "Study drugs were prepared by the hospital pharmacy using a randomisation schedule; all other personnel were blinded to the randomisation. The researchers were informed about the contents of the blinded syringes after the study".
Allocation concealment (selection bias)Low riskQuote: "Study drugs were prepared by the hospital pharmacy using a randomisation schedule; all other personnel were blinded to the randomisation. The researchers were informed about the contents of the blinded syringes after the study".
Blinding of participants and personnel (performance bias)
All outcomes
Low riskQuote: "Study drugs were prepared by the hospital pharmacy using a randomisation schedule; all other personnel were blinded to the randomisation. The researchers were informed about the contents of the blinded syringes after the study".
Blinding of outcome assessment (detection bias)
All outcomes
Low riskQuote: "Study drugs were prepared by the hospital pharmacy using a randomisation schedule; all other personnel were blinded to the randomisation. The researchers were informed about the contents of the blinded syringes after the study".
Incomplete outcome data (attrition bias)
All outcomes
Unclear riskComment: This information was not available.
Selective reporting (reporting bias)High riskComment: Important outcomes were not reported.
Vested interest biasUnclear riskComment: This information was not available.

Garcia-Huete 1997

MethodsRandomised clinical trial.
ParticipantsCountry: Spain.
Sample size: 80.
Post-randomisation drop-out(s): not stated.
Revised sample size: 80.
Females: 29 (36.3%).
Mean age: 50 years.
Piggyback: not stated.
Conventional: not stated.
Cadaveric donor: not stated.
Live donor: not stated.
Inclusion criteria:
Primary elective liver transplantation.
Exclusion criteria:
1. Acute hepatic failure.
2. Retransplantation.
3. Multiorgan transplantation.
InterventionsThe patients were randomised to the following groups.
Group 1: intervention (n = 39).
Further details: aprotinin 2 million KIU at induction followed by a continuous infusion of 500,000 KIU during procedure.
Group 2: control (n = 41).
Further details: placebo (normal saline).
OutcomesThe outcomes reported were mortality, retransplantation, reoperation due to bleeding, and transfusion requirements.
NotesAttempts were made to contact the author in September 2011.
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Unclear riskComment: This information was not available.
Allocation concealment (selection bias)Unclear riskComment: This information was not available.
Blinding of participants and personnel (performance bias)
All outcomes
Low riskQuote: "prospective, randomised, double-blind, and placebo-controlled trial".
Comment: The authors used placebo to achieve the blinding.
Blinding of outcome assessment (detection bias)
All outcomes
Low riskQuote: "prospective, randomised, double-blind, and placebo-controlled trial".
Comment: The authors used placebo to achieve the blinding.
Incomplete outcome data (attrition bias)
All outcomes
Unclear riskComment: This information was not available.
Selective reporting (reporting bias)High riskComment: Important outcomes were not reported.
Vested interest biasUnclear riskComment: This information was not available.

Groh 1993

MethodsRandomised clinical trial.
ParticipantsCountry: Germany.
Sample size: 20.
Post-randomisation drop-out(s): 2 (10%).
Revised sample size: 18.
Females: not stated.
Mean age: 50 years.
Piggyback: not stated.
Conventional: not stated.
Cadaveric donor: not stated.
Live donor: not stated.
Inclusion criteria:
1. Patients undergoing primary elective orthotopic liver transplantation.
InterventionsThe patients were randomised to the following groups.
Group 1: intervention (n = 9).
Further details: aprotinin 2 million KIU loading dose after induction of anaesthesia, followed by continuous infusion of 500,000 KIU/ hour until the end of procedure.
Group 2: control (n = 9).
Further details: placebo.
OutcomesThe outcomes reported were blood transfusion requirements.
NotesAttempts were made to contact the authors in September 2011.

Reason for post-randomisation drop-out(s): death, massive surgical haemorrhage (group not stated).
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Unclear riskComment: This information was not available.
Allocation concealment (selection bias)Unclear riskComment: This information was not available.
Blinding of participants and personnel (performance bias)
All outcomes
Unclear riskComment: Although a placebo was used, the nature of the placebo was not stated.
Blinding of outcome assessment (detection bias)
All outcomes
Unclear riskComment: Although a placebo was used, the nature of the placebo was not stated.
Incomplete outcome data (attrition bias)
All outcomes
High riskQuote: "Two patients had to be excluded from statistical evaluation because of massive surgical haemorrhage and death during the study period".
Selective reporting (reporting bias)High riskComment: Important outcomes were not reported.
Vested interest biasUnclear riskComment: This information was not available.

Hei 2005

MethodsRandomised clinical trial.
ParticipantsCountry: China.
Sample size: 40.
Post-randomisation drop-out(s): not stated.
Revised sample size: 40.
Females: not stated.
Mean age: not stated.
Piggyback: not stated.
Conventional: not stated.
Cadaveric donor: not stated.
Live donor: not stated.
Inclusion criteria:
Patients undergoing liver transplantation.
InterventionsThe patients were randomised to the following groups.
Group 1: intervention (n = 20).
Further details: aprotinin 40,000 KIU/kg/hour continuous infusion.
Group 2: control (n = 20).
Further details: control (no intervention).
OutcomesThe outcomes reported were blood loss and blood transfusion requirements.
NotesAttempts were made to contact the author in September 2011.
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Unclear riskComment: This information was not available.
Allocation concealment (selection bias)Unclear riskComment: This information was not available.
Blinding of participants and personnel (performance bias)
All outcomes
Unclear riskComment: This information was not available.
Blinding of outcome assessment (detection bias)
All outcomes
Unclear riskComment: This information was not available.
Incomplete outcome data (attrition bias)
All outcomes
Unclear riskComment: This information was not available.
Selective reporting (reporting bias)High riskComment: Important outcomes were not reported.
Vested interest biasUnclear riskComment: This information was not available.

Himmelreich 1992

MethodsRandomised clinical trial.
ParticipantsCountry: Germany.
Sample size: 23.
Post-randomisation drop-out(s): not stated.
Revised sample size: 23.
Females: not stated.
Mean age: not stated.
Piggyback: not stated.
Conventional: not stated.
Cadaveric donor: not stated.
Live donor: not stated.
Inclusion criteria:
1. Patients undergoing primary orthotopic liver transplantation.
InterventionsThe patients were randomised to the following groups.
Group 1: intervention (n = 13).
Further details: aprotinin intravenous bolus 0.5 million KIU thrice - at the beginning of the operation, anhepatic phase, and at reperfusion.
Group 2: control (n = 10).
Further details: aprotinin intravenous continuously at 0.2 million KIU per hour from beginning of surgery, increased to 0.4 million KIU per hours during the anhepatic phase and at reperfusion; and decreased to 0.1 million KIU till skin closure.
OutcomesThe outcomes reported were mortality, retransplantation, and transfusion requirements.
NotesAttempts were made to contact the author in September 2011.
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Unclear riskComment: This information was not available.
Allocation concealment (selection bias)Unclear riskQuote: "The patients were randomised by sealed envelopes with serial numbers containing the  way of aprotinin administration".
Comment: Further details of the sealed envelope method was not available
Blinding of participants and personnel (performance bias)
All outcomes
Unclear riskComment: This information was not available.
Blinding of outcome assessment (detection bias)
All outcomes
Unclear riskComment: This information was not available.
Incomplete outcome data (attrition bias)
All outcomes
Unclear riskComment: This information was not available.
Selective reporting (reporting bias)High riskComment: Important outcomes were not reported.
Vested interest biasUnclear riskComment: This information was not available.

Himmelreich 1993

MethodsRandomised clinical trial.
ParticipantsCountry: Germany.
Sample size: 20.
Post-randomisation drop-out(s): not stated.
Revised sample size: 20.
Females: not stated.
Mean age: not stated.
Piggyback: not stated.
Conventional: not stated.
Cadaveric donor: not stated.
Live donor: not stated.
Inclusion criteria:
1. Patients undergoing orthotopic liver transplantation.
InterventionsThe patients were randomised to the following groups.
Group 1: intervention (n = 10).
Further details: prostaglandin (PGE1) at 10 mcg/hour increased to 40 mcg/hour in steps of 10 mcg/hour from the beginning of the operation; (infusion was temporarily stopped if the systolic pressure was less than 100 mm hg.
Group 2: control (n = 10).
Further details: control (no intervention).
OutcomesThe outcomes reported were retransplantation and blood transfusion requirements.
NotesAttempts were made to contact the author in September 2011.
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Unclear riskComment: This information was not available.
Allocation concealment (selection bias)Unclear riskComment: This information was not available.
Blinding of participants and personnel (performance bias)
All outcomes
Unclear riskComment: This information was not available.
Blinding of outcome assessment (detection bias)
All outcomes
Unclear riskComment: This information was not available.
Incomplete outcome data (attrition bias)
All outcomes
Low riskQuote: "In a consecutive series, twenty patients with final liver disease underwent their first OLT at the University Hospital Rudolf Virchow, Berlin, between August and November 1990".
Selective reporting (reporting bias)High riskComment: Important outcomes were not reported.
Vested interest biasUnclear riskComment: This information was not available.

Ickx 1993

MethodsRandomised clinical trial.
ParticipantsCountry: Belgium.
Sample size: 19.
Post-randomisation drop-out(s): not stated.
Revised sample size: 19.
Females: not stated.
Mean age: not stated.
Piggyback: not stated.
Conventional: not stated.
Cadaveric donor: not stated.
Live donor: not stated.
Inclusion criteria:
1. Patients undergoing orthotopic liver transplantation.
InterventionsThe patients were randomised to the following groups.
Group 1: intervention (n = 10).
Further details: aprotinin 187,500 KIU/hour from the beginning of the procedure.
Group 2: control (n = 9).
Further details: control (placebo).
OutcomesThe outcomes reported were blood loss and blood transfusion requirements.
NotesAttempts were made to contact the author in September 2011.
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Unclear riskComment: This information was not available.
Allocation concealment (selection bias)Unclear riskComment: This information was not available.
Blinding of participants and personnel (performance bias)
All outcomes
Unclear riskComment: A placebo was used. It was not clear what was used for placebo.
Blinding of outcome assessment (detection bias)
All outcomes
Unclear riskComment: A placebo was used. It was not clear what was used for placebo.
Incomplete outcome data (attrition bias)
All outcomes
Unclear riskComment: This information was not available.
Selective reporting (reporting bias)High riskComment: Important outcomes were not reported.
Vested interest biasUnclear riskComment: This information was not available.

Ickx 1995

MethodsRandomised clinical trial.
ParticipantsCountry: Belgium.
Sample size: 20.
Post-randomisation drop-out(s): not stated.
Revised sample size: 20.
Females: not stated.
Mean age: not stated.
Piggyback: not stated.
Conventional: not stated.
Cadaveric donor: not stated.
Live donor: not stated.
Inclusion criteria:
1. Patients undergoing orthotopic liver transplantation.
InterventionsThe patients were randomised to the following groups.
Group 1: intervention (n = 10).
Further details: tranexamic acid 80 mg/kg bolus followed by 40 mg/kg/hour.
Group 2: control (n = 10).
Further details: aprotinin 2 million KIU bolus followed by 0.5 million KIU/hour.
OutcomesThe outcomes reported were blood loss and blood transfusion requirements.
NotesAttempts were made to contact the author in September 2011.
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Unclear riskComment: This information was not available.
Allocation concealment (selection bias)Unclear riskComment: This information was not available.
Blinding of participants and personnel (performance bias)
All outcomes
Unclear riskComment: This information was not available.
Blinding of outcome assessment (detection bias)
All outcomes
Unclear riskComment: This information was not available.
Incomplete outcome data (attrition bias)
All outcomes
Unclear riskComment: This information was not available.
Selective reporting (reporting bias)High riskComment: Important outcomes were not reported.
Vested interest biasUnclear riskComment: This information was not available.

Ickx 2006

MethodsRandomised clinical trial.
ParticipantsCountry: Belgium.
Sample size: 51.
Post-randomisation drop-out(s): not stated.
Revised sample size: 51.
Females: not stated.
Mean age: not stated.
Piggyback: not stated.
Conventional: not stated.
Cadaveric donor: not stated.
Live donor: not stated.
Inclusion criteria:
1. Patients undergoing primary orthotopic liver transplantation.
Exclusion criteria:
1. Fulminant hepatitis.
2. Past history of thromboembolic events.
3. Retransplantation.
Interventions

The patients were randomised to the following groups.
Group 1: intervention (n = 27).
Further details: tranexamic acid 40 mg/kg as bolus initiated during the anhepatic phase and maintained at 40 mg/kg/hour as continuous infusion until 2 hours after reperfusion.

Group 2: control (n = 24).
Further details: aprotinin 2 million KIU as bolus initiated during the anhepatic phase and maintained at 0.5 million KIU as continuous infusion until 2 hours after reperfusion.

OutcomesThe outcomes reported were mortality, hospital stay, blood loss, and blood transfusion requirements.
NotesAttempts were made to contact the author in September 2011.
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Unclear riskComment: This information was not available.
Allocation concealment (selection bias)Unclear riskComment: This information was not available.
Blinding of participants and personnel (performance bias)
All outcomes
Unclear riskComment: This information was not available.
Blinding of outcome assessment (detection bias)
All outcomes
Unclear riskComment: This information was not available.
Incomplete outcome data (attrition bias)
All outcomes
Unclear riskComment: This information was not available.
Selective reporting (reporting bias)High riskComment: Important outcomes were not reported.
Vested interest biasHigh riskQuote: "Drugs were provided by Choay, Bournonville, France".

Kaspar 1997

MethodsRandomised clinical trial.
ParticipantsCountry: USA.
Sample size: 32.
Post-randomisation drop-out(s): not stated.
Revised sample size: 32.
Females: not stated.
Mean age: not stated.
Piggyback: not stated.
Conventional: not stated.
Cadaveric donor: not stated.
Live donor: not stated.
Inclusion criteria:
1. Patients undergoing orthotopic liver transplantation.
InterventionsThe patients were randomised to the following groups.
Group 1: intervention (n = 16).
Further details: tranexamic acid 2 mg/kg/hour after induction of anaesthesia until the end of surgery.
Group 2: control (n = 16).
Further details: control (placebo - normal saline).
OutcomesThe outcomes reported were mortality, primary non-function, retransplantation, and blood transfusion requirements.
NotesAttempts were made to contact the author in September 2011.
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Low riskQuote: "The infusions were prepared by the hospital pharmacy using a computer generated randomisation schedule. All investigators were blinded to the composition of the solutions".
Allocation concealment (selection bias)Low riskQuote: "The infusions were prepared by the hospital pharmacy using a computer generated randomisation schedule. All investigators were blinded to the composition of the solutions".
Blinding of participants and personnel (performance bias)
All outcomes
Low riskQuote: "The infusions were prepared by the hospital pharmacy using a computer generated randomisation schedule. All investigators were blinded to the composition of the solutions".
Blinding of outcome assessment (detection bias)
All outcomes
Low riskQuote: "The infusions were prepared by the hospital pharmacy using a computer generated randomisation schedule. All investigators were blinded to the composition of the solutions".
Incomplete outcome data (attrition bias)
All outcomes
Low riskQuote: "After institutional review board approval and informed consent, a prospective, placebo-controlled, double-blind study was performed on 32 consecutive patients undergoing OLT".
Selective reporting (reporting bias)High riskComment: Important outcomes were not reported.
Vested interest biasUnclear riskComment: This information was not available.

Laine 2003

MethodsRandomised clinical trial.
ParticipantsCountry: USA.
Sample size: 33.
Post-randomisation drop-out(s): not stated.
Revised sample size: 33.
Females: not stated.
Mean age: not stated.
Piggyback: not stated.
Conventional: not stated.
Cadaveric donor: not stated.
Live donor: not stated.
Inclusion criteria:
1. Patients undergoing liver transplantation.
2. Age more than 18 years.
Exclusion criteria:
1. Patients with B or AB negative blood group.
2. Patients with serum alloantibodies.
3. ABO mismatched donor liver.
InterventionsThe patients were randomised to the following groups.
Group 1: intervention (n = 18).
Further details: whole blood.
Group 2: control (n = 15).
Further details: component therapy.
OutcomesThe outcomes reported were volume  of fluid transfused. None of the outcomes included in this review were reported in this trial.
NotesAttempts were made to contact the authors in September 2011.
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Unclear riskComment: This information was not available.
Allocation concealment (selection bias)Unclear riskComment: This information was not available.
Blinding of participants and personnel (performance bias)
All outcomes
Unclear riskComment: This information was not available.
Blinding of outcome assessment (detection bias)
All outcomes
Unclear riskComment: This information was not available.
Incomplete outcome data (attrition bias)
All outcomes
Unclear riskComment: This information was not available.
Selective reporting (reporting bias)High riskComment: Important outcomes were not reported.
Vested interest biasUnclear riskComment: This information was not available.

Lassale 1996

MethodsRandomised clinical trial.
ParticipantsCountry: France.
Sample size: 20.
Post-randomisation drop-out(s): not stated.
Revised sample size: 20.
Females: 7 (35%).
Mean age: 47.5 years.
Piggyback: not stated.
Conventional: not stated.
Cadaveric donor: not stated.
Live donor: not stated.
Inclusion criteria:
1. Patients undergoing primary liver transplantation for chronic indications.
2. Age more than 18 years.
Exclusion criteria:
1. Retransplantation.
2. Allergies to aprotinin.
3. Prior acute pancreatitis.
InterventionsThe patients were randomised to the following groups.
Group 1: intervention (n = 10).
Further details: aprotinin 100,000 KIU/hour by continuous infusion.
Group 2: control (n = 10).
Further details: aprotinin 200,000 KIU/hour by continuous infusion.
OutcomesThe outcomes reported were blood transfusion requirements.
NotesAttempts were made to contact the author in September 2011.
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Low riskQuote: "Le protocole d’administration a été attribué au hasard par une table de randomisation".
Allocation concealment (selection bias)Unclear riskComment: This information was not available.
Blinding of participants and personnel (performance bias)
All outcomes
Unclear riskComment: This information was not available.
Blinding of outcome assessment (detection bias)
All outcomes
Unclear riskComment: This information was not available.
Incomplete outcome data (attrition bias)
All outcomes
Unclear riskComment: This information was not available.
Selective reporting (reporting bias)High riskComment: Important outcomes were not reported.
Vested interest biasUnclear riskComment: This information was not available.

Lodge 2005

MethodsRandomised clinical trial.
ParticipantsCountry: Australia, Canada, Denmark, Germany, Spain, Sweden, United Kingdom.
Sample size: 209.
Post-randomisation drop-out(s): 26 (12.4%).
Revised sample size: 183.
Females: 72 (34.4%).
Mean age: 52.7 years.
Piggyback: not stated.
Conventional: not stated.
Cadaveric donor: 182 (87.1%).
Live donor: 0 (0%).
Inclusion criteria:
1. Patients scheduled to undergo orthotopic liver transplantation.
2. Age greater than 18 years.
Exclusion criteria:
1. Previous liver transplantation.
2. Split liver transplantation.
3. Multiorgan transplantation.
4.Living related-donor transplantation.
5. Renal insufficiency requiring dialysis.
6. Documented coagulation disorders.
7. Documented history or presence of portal vein thrombosis.
InterventionsThe patients were randomised to the following groups.
Group 1: intervention 1 (n = 58).
Further details of intervention: rFVIIa 120 mcg/Kg as bolus doses starting from within 10 minutes of skin incision; repeated every 2 hours until 30 minutes before expected reperfusion; and a final dose on skin closure.
Group 2: intervention 2 (n = 63).
Further details of intervention: rFVIIa 60 mcg/Kg as bolus doses starting from within 10 minutes of skin incision; repeated every 2 hours until 30 minutes before expected reperfusion; and a final dose on skin closure.
Group 3: control (n = 61).
Further details of intervention: placebo (not specified).
OutcomesThe outcomes reported were mortality, adverse events, hospital stay and transfusion requirements.
NotesAttempts were made to contact the author in September 2011.

Reason for post-randomisation drop-out(s): Did not meet the inclusion criteria on the day of operation.
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Unclear riskComment: This information was not available.
Allocation concealment (selection bias)Unclear riskComment: This information was not available.
Blinding of participants and personnel (performance bias)
All outcomes
Unclear riskComment: Although the authors state it is a double blinded placebo controlled randomised trial, the nature of the placebo was not known.
Blinding of outcome assessment (detection bias)
All outcomes
Unclear riskComment: Although the authors state it is a double blinded placebo controlled randomised trial, the nature of the placebo was not known.
Incomplete outcome data (attrition bias)
All outcomes
High riskComment: There were post-randomisation drop-outs.
Selective reporting (reporting bias)High riskComment: Important outcomes were not reported.
Vested interest biasUnclear riskComment: This information was not available.

Marcel 1996

MethodsRandomised clinical trial.
ParticipantsCountry: USA.
Sample size: 44.
Post-randomisation drop-out(s): not stated.
Revised sample size: 44.
Females: not stated.
Mean age: not stated.
Piggyback: not stated.
Conventional: not stated.
Cadaveric donor: not stated.
Live donor: not stated.
Inclusion criteria:
1. Patients undergoing orthotopic liver transplantation with venovenous bypass.
InterventionsThe patients were randomised to the following groups.
Group 1: intervention (n = 21).
Further details: aprotinin 200,000 KIU/hour immediately after induction (duration not stated).
Group 2: control (n = 23).
Further details: placebo (normal saline).
OutcomesThe outcomes reported were blood transfusion requirements.
NotesAttempts were made to contact the author in September 2011.
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Low riskQuote: "Participants were randomised by a computer program to receive, via an infusion pump, identical volumes of aprotinin (n = 21) or normal saline solution (n = 23)".
Allocation concealment (selection bias)Unclear riskComment: This information was not available.
Blinding of participants and personnel (performance bias)
All outcomes
Low riskQuote: "After institutional review board approval and informed consent, a prospective, placebo-controlled, double-blind study was performed on 44 consecutive patients undergoing OLT with veno-venous bypass".
Comment: Normal saline was used as placebo.
Blinding of outcome assessment (detection bias)
All outcomes
Low riskQuote: "After institutional review board approval and informed consent, a prospective, placebo-controlled, double-blind study was performed on 44 consecutive patients undergoing OLT with veno-venous bypass".
Comment: Normal saline was used as placebo.
Incomplete outcome data (attrition bias)
All outcomes
Low riskComment: There were no post-randomisation drop-outs.
Selective reporting (reporting bias)High riskComment: Important outcomes were not reported.
Vested interest biasUnclear riskComment: This information was not available.

Merle 1997

MethodsRandomised clinical trial.
ParticipantsCountry: Netherlands.
Sample size: 20.
Post-randomisation drop-out(s): not stated.
Revised sample size: 20.
Females: not stated.
Mean age: not stated.
Piggyback: not stated.
Conventional: not stated.
Cadaveric donor: not stated.
Live donor: not stated.
Inclusion criteria:
1. Cirrhotic patients with severe liver failure undergoing liver transplantation.
InterventionsThe patients were randomised to the following groups.
Group 1: intervention (n = 10).
Further details: aprotinin 2 million KIU prior to skin incision followed by continuous infusion of 0.5 million KIU until the end of surgery.
Group 2: control (n = 10).
Further details: placebo (nature not stated).
OutcomesThe outcomes reported were blood transfusion requirements.
NotesAttempts were made to contact the authors in September 2011.
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Unclear riskComment: This information was not available.
Allocation concealment (selection bias)Unclear riskComment: This information was not available.
Blinding of participants and personnel (performance bias)
All outcomes
Unclear riskComment: Although the authors state that they used placebo, the nature of the placebo was not stated.
Blinding of outcome assessment (detection bias)
All outcomes
Unclear riskComment: Although the authors state that they used placebo, the nature of the placebo was not stated.
Incomplete outcome data (attrition bias)
All outcomes
Low riskComment: There were no post-randomisation drop-outs.
Selective reporting (reporting bias)High riskComment: Important outcomes were not reported.
Vested interest biasUnclear riskComment: This information was not available.

Milroy 1995

MethodsRandomised clinical trial.
ParticipantsCountry: United Kingdom.
Sample size: 55.
Post-randomisation drop-out(s): 3 (5.5%).
Revised sample size: 52.
Females: 26 (47.3%).
Mean age: 44.8 years.
Piggyback: not stated.
Conventional: not stated.
Cadaveric donor: not stated.
Live donor: not stated.
Inclusion criteria:
1. Patients with chronic or fulminant liver failure undergoing orthotopic liver transplantation.
2. Age greater than 18 years.
Exclusion criteria:
1. Previous exposure or allergy to aprotinin.
2. History of pancreatitis.
3. Oliguric renal failure.
InterventionsThe patients were randomised to the following groups.
Group 1: intervention (n = 26).
Further details: aprotinin 2 million KIU bolus after induction before laparotomy followed by continuous infusion at 0.5 million KIU/hour until the patient was transferred to ITU.
Group 2: control (n = 26).
Further details: placebo (normal saline).
OutcomesThe outcomes reported were haemodynamic parameters. None of the outcomes of interest for this review were reported.
NotesAttempts were made to contact the authors in September 2011.

Reason for post-randomisation drop-out(s): patient died because of heart failure (1); operation abandoned due to tissue oedema (1); inoperable liver tumour (1).
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Low riskQuote: "Randomization was performed by a previously determined computer-generated schedule using coded ampoules which were supplied by the manufacturer (Bayer plc) in identical case packs. The investigators were unaware of the coding schedule".
Allocation concealment (selection bias)Low riskQuote: "Randomization was performed by a previously determined computer-generated schedule using coded ampoules which were supplied by the manufacturer (Bayer plc) in identical case packs. The investigators were unaware of the coding schedule".
Blinding of participants and personnel (performance bias)
All outcomes
Low riskQuote: "Randomization was performed by a previously determined computer-generated schedule using coded ampoules which were supplied by the manufacturer (Bayer plc) in identical case packs. The investigators were unaware of the coding schedule".
Blinding of outcome assessment (detection bias)
All outcomes
Low riskQuote: "Randomization was performed by a previously determined computer-generated schedule using coded ampoules which were supplied by the manufacturer (Bayer plc) in identical case packs. The investigators were unaware of the coding schedule".
Incomplete outcome data (attrition bias)
All outcomes
High riskComment: There were post-randomisation drop-outs.
Selective reporting (reporting bias)High riskComment: Important outcomes were not reported.
Vested interest biasHigh riskQuote: "We thank Bayer Pharmaceuticals  for their support and sponsorship".

Planinsic 2005

MethodsRandomised clinical trial.
ParticipantsCountry: Denmark, Finland, Germany, Spain, USA.
Sample size: 87.
Post-randomisation drop-out(s): 4 (4.6%).
Revised sample size: 83.
Females: 26 (29.9%).
Mean age: 50.2 years.
Piggyback: not stated.
Conventional: not stated.
Cadaveric donor: not stated.
Live donor: not stated.
Inclusion criteria:
1. Patients scheduled to undergo orthotopic liver transplantation.
2. Age more than 18 years.
Exclusion criteria:
1. Previous liver transplantation.
2. Multiorgan transplantation.
3. Living related-donor transplantation.
4. Renal insufficiency requiring dialysis.
5. Documented inherited coagulation disorders.
6. Documented history or presence of portal vein thrombosis.
InterventionsThe patients were randomised to the following groups.
Group 1: intervention 1 (n = 22).
Further details of intervention: rFVIIa 80 mcg/Kg as bolus dose within 10 minutes of skin incision.
Group 2: intervention 2 (n = 24).
Further details of intervention: rFVIIa 40 mcg/Kg as bolus dose within 10 minutes of skin incision.
Group 3: intervention 2 (n = 18).
Further details of intervention: rFVIIa 20 mcg/Kg as bolus dose within 10 minutes of skin incision.
Group 3: control (n = 19).
Further details of intervention: placebo (not specified).
OutcomesThe outcomes reported were serious adverse events and blood transfusion requirements.
NotesAttempts were made to contact the author in September 2011.

Reason for post-randomisation drop-out(s): liver transplantation cancelled.
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Unclear riskComment: This information was not available.
Allocation concealment (selection bias)Unclear riskComment: This information was not available.
Blinding of participants and personnel (performance bias)
All outcomes
Unclear riskComment: Although the authors state placebo was used, the nature of the placebo was not stated.
Blinding of outcome assessment (detection bias)
All outcomes
Unclear riskComment: Although the authors state placebo was used, the nature of the placebo was not stated.
Incomplete outcome data (attrition bias)
All outcomes
High riskComment: There were post-randomisation drop-outs.
Selective reporting (reporting bias)High riskComment: Important outcomes were not reported.
Vested interest biasUnclear riskComment: This information was not available.

Ponnudurai 2005

MethodsRandomised clinical trial.
ParticipantsCountry: USA.
Sample size: 65.
Post-randomisation drop-out(s): not stated.
Revised sample size: 65.
Females: 21 (32.3%).
Mean age: 53.4 years.
Piggyback: not stated.
Conventional: not stated.
Cadaveric donor: not stated.
Live donor: not stated.
Inclusion criteria:
1. Adult patients about to undergo orthotopic liver transplantation for endstage liver disease.
Exclusion criteria:
1. Patients undergoing combined liver and kidney transplantation.
2. Retransplantation.
InterventionsThe patients were randomised to the following groups.
Group 1: intervention (n = 33).
Further details: norepinephrine 0.5 mcg/min increased in 1 mcg increments up to 6 mcg/min to maintain a mean arterial pressure and pulmonary capillary wedge pressures at > 80% of baseline values.
Group 2: control (n = 32).
Further details: placebo (normal saline).
OutcomesThe outcomes reported were hospital stay and blood transfusion requirements.
NotesAttempts were made to contact the author in September 2011.
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Low riskQuote: "Using a computer-generated list of random numbers, patients were assigned to receive norepinephrine (vasopressor group) or IV fluids and isotonic saline (placebo group) by the research coordinator, who had no part in the clinical management of patients".
Allocation concealment (selection bias)Low riskQuote: "Using a computer-generated list of random numbers, patients were assigned to receive norepinephrine (vasopressor group) or IV fluids and isotonic saline (placebo group) by the research coordinator, who had no part in the clinical management of patients".
Blinding of participants and personnel (performance bias)
All outcomes
Low riskQuote: "The blinding of the investigators to isotonic saline or norepinephrine was done by the Department of Pharmacy at the UMDNJ. Blinding was maintained throughout the study period by pharmacists, who had no role in the clinical management of the study patients and data collection. The vasopressor and placebo were both colorless fluids".
Blinding of outcome assessment (detection bias)
All outcomes
Low riskQuote: "The blinding of the investigators to isotonic saline or norepinephrine was done by the Department of Pharmacy at the UMDNJ. Blinding was maintained throughout the study period by pharmacists, who had no role in the clinical management of the study patients and data collection. The vasopressor and placebo were both colorless fluids".
Incomplete outcome data (attrition bias)
All outcomes
Unclear riskComment: This information was not available.
Selective reporting (reporting bias)High riskComment: Important outcomes were not reported.
Vested interest biasUnclear riskComment: This information was not available.

Porte 2000

MethodsRandomised clinical trial.
ParticipantsCountry: Finland, Italy, Netherlands, Sweden.
Sample size: 141.
Post-randomisation drop-out(s): 5 (3.5%).
Revised sample size: 136.
Females: 33 (23.4%).
Mean age: 51.4 years.
Piggyback: not stated.
Conventional: not stated.
Cadaveric donor: not stated.
Live donor: not stated.
Inclusion criteria:
1. Patients who underwent primary orthotopic transplantation for acute or chronic liver disease.
2. Age at least 18 years.
Exclusion criteria:
1. Retransplantation.
2. Known or suspected exposure to aprotinin.
3. Malignant disease.
4. pre-existing thrombotic conditions.
InterventionsThe patients were randomised to the following groups.
Group 1: intervention 1 (n = 45).
Further details of intervention: 2 million KIU aprotinin as a loading dose given intravenously over 20 min before and during induction of anaesthesia, followed by a continuous infusion of 1 million KIU/hour until 2 hours after graft reperfusion. An additional dose of one million KIU was given half an hour before reperfusion.
Group 2: intervention 2 (n = 43).
Further details of intervention: 2 million KIU aprotinin as a loading dose given intravenously over 20 min before and during induction of anaesthesia, followed by a continuous infusion of 0.5 million KIU/hour until 2 hours after graft reperfusion.
Group 3: control (n = 48).
Further details: placebo (not stated).
OutcomesThe outcomes reported were mortality, primary graft non-function, retransplantation, reoperation, adverse effects, ITU stay, and blood transfusion requirements.
NotesAttempts were made to contact the authors in September 2011.

Reason for post-randomisation drop-out(s): protocol violation (5); allergic reaction (1).
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Low riskComment: From the other details available in this trial, it is highly likely that the authors used an appropriate method of random sequence generation.
Allocation concealment (selection bias)Low riskQuote: "The trial drug was provided double blind by the manufacturer in blocks of 12 identical case packs. Each case pack contained all bottles for one patient, identifiable only by the sequence number".
Blinding of participants and personnel (performance bias)
All outcomes
Unclear riskQuote: "The trial drug was provided double blind by the manufacturer in blocks of 12 identical case packs. Each case pack contained all bottles for one patient, identifiable only by the sequence number".
Comment: The nature of placebo was not stated.
Blinding of outcome assessment (detection bias)
All outcomes
Unclear riskQuote: "The trial drug was provided double blind by the manufacturer in blocks of 12 identical case packs. Each case pack contained all bottles for one patient, identifiable only by the sequence number".
Comment: The nature of placebo was not stated.
Incomplete outcome data (attrition bias)
All outcomes
High riskComment: There were post-randomisation drop-outs.
Selective reporting (reporting bias)High riskComment: Important outcomes were not reported.
Vested interest biasHigh riskQuote: "Aprotinin and placebo were donated by Bayer, Wuppertal, Germany".

Pugliese 2007

MethodsRandomised clinical trial.
ParticipantsCountry: Italy.
Sample size: 20.
Post-randomisation drop-out(s): not stated.
Revised sample size: 20.
Females: not stated.
Mean age: not stated.
Piggyback: not stated.
Conventional: not stated.
Cadaveric donor: not stated.
Live donor: not stated.
Inclusion criteria:
1. Patients undergoing orthotopic liver transplantation.
InterventionsThe patients were randomised to the following groups.
Group 1: intervention (n = 10).
Further details: rFVIIa 40 mcg/kg as bolus immediately before anaesthesia induction.
Group 2: control (n = 10).
Further details: placebo (details not known).
OutcomesThe outcomes reported were mortality, primary graft non-function, ITU stay, and blood transfusion requirements (the transfusion requirements during the entire operation and immediate post-operative period was not reported and so this information could not be used).
NotesAttempts were made to contact the authors in September 2011.
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Unclear riskComment: This information was not available.
Allocation concealment (selection bias)Unclear riskComment: This information was not available.
Blinding of participants and personnel (performance bias)
All outcomes
Unclear riskComment: Although the authors state that a placebo was used, they do not state the nature of the placebo or the groups who knew whether the test drug was rFVIIa or placebo.
Blinding of outcome assessment (detection bias)
All outcomes
Unclear riskComment: Although the authors state that a placebo was used, they do not state the nature of the placebo or the groups who knew whether the test drug was rFVIIa or placebo.
Incomplete outcome data (attrition bias)
All outcomes
Unclear riskComment: This information was not available.
Selective reporting (reporting bias)High riskComment: Important outcomes were not reported.
Vested interest biasUnclear riskComment: This information was not available.

Rummo 2010

MethodsRandomised clinical trial.
ParticipantsCountry: Belarus.
Sample size: 34.
Post-randomisation drop-out(s): not stated.
Revised sample size: 34.
Females: not stated.
Mean age: 43 years.
Piggyback: not stated.
Conventional: not stated.
Cadaveric donor: 34 (100%).
Live donor: 0 (0%).
Inclusion criteria:
1. Patients undergoing orthotopic liver transplantation.
2. Cadaveric donors.
InterventionsThe patients were randomised to the following groups.
Group 1: intervention (n = 17).
Further details: thromboelastography to determine transfusion requirements.
Group 2: control (n = 17).
Further details: standard method to determine transfusion requirements.
OutcomesThe outcomes reported were sepsis, blood loss, and blood transfusion requirements.
NotesAttempts were made to contact the author in September 2011.
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Unclear riskComment: This information was not available.
Allocation concealment (selection bias)Unclear riskComment: This information was not available.
Blinding of participants and personnel (performance bias)
All outcomes
Unclear riskComment: This information was not available.
Blinding of outcome assessment (detection bias)
All outcomes
Unclear riskComment: This information was not available.
Incomplete outcome data (attrition bias)
All outcomes
Unclear riskComment: This information was not available.
Selective reporting (reporting bias)High riskComment: Important outcomes were not reported.
Vested interest biasUnclear riskComment: This information was not available.

Soilleux 1995

MethodsRandomised clinical trial.
ParticipantsCountry: France.
Sample size: 189.
Post-randomisation drop-out(s): not stated.
Revised sample size: 189.
Females: not stated.
Mean age: 43.5 years.
Piggyback: not stated.
Conventional: not stated.
Cadaveric donor: not stated.
Live donor: not stated.
Inclusion criteria:
1. Adult patients undergoing primary orthotopic liver transplantation.
Exclusion criteria:
1. Retransplantation
InterventionsThe patients were randomised to the following groups.
Group 1: intervention (n = 94).
Further details: aprotinin 2 million KIU bolus on induction followed by continuous infusion of 0.5 million KIU until transfer to ITU.
Group 2: control (n = 95).
Further details: aprotinin 0.5 million KIU bolus on induction followed by continuous infusion of 150,000 KIU until transfer to ITU.
OutcomesThe outcomes reported were blood transfusion requirements.
NotesAttempts were made to contact the authors in September 2011.
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Low riskQuote: "Using a computer-generated list of random numbers, patients were assigned to receive norepinephrine (vasopressor group) or IV fluids and isotonic saline (placebo group) by the research coordinator, who had no part in the clinical management of patients".
Allocation concealment (selection bias)Low riskQuote: "Using a computer-generated list of random numbers, patients were assigned to receive norepinephrine (vasopressor group) or IV fluids and isotonic saline (placebo group) by the research coordinator, who had no part in the clinical management of patients".
Blinding of participants and personnel (performance bias)
All outcomes
Low riskQuote: "The blinding of the investigators to isotonic saline or norepinephrine was done by the Department of Pharmacy at the UMDNJ. Blinding was maintained throughout the study period by pharmacists, who had no role in the clinical management of the study patients and data collection. The vasopressor and placebo were both colorless fluids".
Blinding of outcome assessment (detection bias)
All outcomes
Low riskQuote: "The blinding of the investigators to isotonic saline or norepinephrine was done by the Department of Pharmacy at the UMDNJ. Blinding was maintained throughout the study period by pharmacists, who had no role in the clinical management of the study patients and data collection. The vasopressor and placebo were both colorless fluids".
Incomplete outcome data (attrition bias)
All outcomes
Unclear riskComment: This information was not available.
Selective reporting (reporting bias)High riskComment: Important outcomes were not reported.
Vested interest biasUnclear riskComment: This information was not available.

Wang 2010

MethodsRandomised clinical trial.
ParticipantsCountry: Taiwan.
Sample size: 28.
Post-randomisation drop-out(s): not stated.
Revised sample size: 28.
Females: 10 (35.7%).
Mean age: 55 years.
Piggyback: not stated.
Conventional: not stated.
Cadaveric donor: not stated.
Live donor: not stated.
Inclusion criteria:
1. Patients undergoing orthotopic liver transplantation.
InterventionsThe patients were randomised to the following groups.
Group 1: intervention (n = 14).
Further details: thromboelastography determined decision to transfuse.
Group 2: control (n = 14).
Further details: standard method to determine decision to transfuse.
OutcomesThe outcomes reported were mortality, blood loss, and transfusion requirements.
NotesAttempts were made to contact the author in September 2011.
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Unclear riskComment: This information was not available.
Allocation concealment (selection bias)Unclear riskComment: This information was not available.
Blinding of participants and personnel (performance bias)
All outcomes
Unclear riskComment: This information was not available.
Blinding of outcome assessment (detection bias)
All outcomes
Unclear riskComment: This information was not available.
Incomplete outcome data (attrition bias)
All outcomes
Unclear riskComment: This information was not available.
Selective reporting (reporting bias)High riskComment: Important outcomes were not reported.
Vested interest biasUnclear riskComment: This information was not available.

Williamson 1999

MethodsRandomised clinical trial.
ParticipantsCountry: United Kingdom.
Sample size: 28.
Post-randomisation drop-out(s): 3 (10.7%).
Revised sample size: 25.
Females: 13 (46.4%).
Mean age: 49 years.
Piggyback: not stated.
Conventional: not stated.
Cadaveric donor: not stated.
Live donor: not stated.
Inclusion criteria:
1. Adult patients with liver disease or undergoing liver transplantation (only liver transplantation patients were included in this review).
InterventionsThe patients were randomised to the following groups.
Group 1: intervention (n = 12).
Further details: solvent detergent treated fresh frozen plasma.
Group 2: control (n = 13).
Further details: standard fresh frozen plasma.
OutcomesThe outcomes reported were blood transfusion requirements.
NotesAttempts were made to contact the author in September 2011.

Reason for post-randomisation drop-out(s): protocol violation (3).
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Low riskQuote: "Assignment was accomplished by the opening of a previously allocated, computer-generated, randomly numbered envelope".
Allocation concealment (selection bias)Unclear risk

Quote: "Assignment was accomplished by the opening of a previously allocated, computer-generated, randomly numbered envelope".

Comment: Further details were not available.

Blinding of participants and personnel (performance bias)
All outcomes
Unclear riskComment: This information was not available.
Blinding of outcome assessment (detection bias)
All outcomes
Unclear riskComment: This information was not available.
Incomplete outcome data (attrition bias)
All outcomes
High riskComment: There were post-randomisation drop-outs.
Selective reporting (reporting bias)High riskComment: Important outcomes were not reported.
Vested interest biasHigh riskQuote: "Supported in part by Octapharma AG (Vienna, Austria)".

Yassen 1993

MethodsRandomised clinical trial.
ParticipantsCountry: United Kingdom.
Sample size: 20.
Post-randomisation drop-out(s): not stated.
Revised sample size: 20.
Females: 11 (55%).
Mean age: 47.2 years.
Piggyback: not stated.
Conventional: not stated.
Cadaveric donor: not stated.
Live donor: not stated.
Inclusion criteria:
1. Patients undergoing orthotopic liver transplantation.
InterventionsThe patients were randomised to the following groups.
Group 1: intervention (n = 10).
Further details: tranexamic acid 10 mg/kg bolus dose at the start of anhepatic phase followed by 3 mg/kg/hour until the patient is transferred to ITU.
Group 2: control (n = 10).
OutcomesThe outcomes reported were serious adverse effects, blood loss, and transfusion requirements.
NotesAttempts were made to contact the author in September 2011.
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Unclear riskComment: This information was not available.
Allocation concealment (selection bias)Unclear riskComment: This information was not available.
Blinding of participants and personnel (performance bias)
All outcomes
Unclear riskComment: This information was not available.
Blinding of outcome assessment (detection bias)
All outcomes
Unclear riskComment: This information was not available.
Incomplete outcome data (attrition bias)
All outcomes
Unclear riskComment: This information was not available.
Selective reporting (reporting bias)High riskComment: Important outcomes were not reported.
Vested interest biasUnclear riskComment: This information was not available.

Characteristics of excluded studies [ordered by study ID]

StudyReason for exclusion
  1. a

    RCT = randomised clinical trial.

Chapman 2006Does not include patients undergoing liver transplantation.
Grosse 1991Not a randomised clinical trial.
Grosse 1993Not a randomised clinical trial.
Ickx 1993aNot a randomised clinical trial.
Kang 1993Review.
Kufner 2000Review.
Lentschener 1996Letter to editor with reference to included trial.
Levy 2008Review.
Marino 1988Not a randomised clinical trial.
Meijer 2003Not a randomised clinical trial.
Molenaar 2002Letter to editor with no data from RCT.
Neuhaus 1989Not a randomised clinical trial.
Palareti 1991Not a randomised clinical trial.
Porte 2001Editorial.
Porte 2004Editorial.
Porte 2004aReview.
Porte 2005Editorial.
Rosea 2000Editorial.
Sankarankutty 2006Not a randomised clinical trial.
Segal 2000Letter to editor with no data from RCT.
Spohr 2002Review.
Suárez 1993Not a randomised clinical trial.
Takaya 1995Not a randomised clinical trial.

Ancillary