Intervention Review

You have free access to this content

Antithrombin III for critically ill patients

  1. Arash Afshari1,*,
  2. Jørn Wetterslev2,
  3. Jesper Brok3,
  4. Ann Merete Møller4

Editorial Group: Cochrane Anaesthesia Group

Published Online: 16 JUL 2008

Assessed as up-to-date: 4 NOV 2006

DOI: 10.1002/14651858.CD005370.pub2


How to Cite

Afshari A, Wetterslev J, Brok J, Møller AM. Antithrombin III for critically ill patients. Cochrane Database of Systematic Reviews 2008, Issue 3. Art. No.: CD005370. DOI: 10.1002/14651858.CD005370.pub2.

Author Information

  1. 1

    Rigshospitalet, The Cochrane Anaesthesia Review Group & Copenhagen Trial Unit and Department of Paediatric and Obstetric Anaesthesia, Copenhagen, Denmark

  2. 2

    Rigshospitalet, Copenhagen University Hospital, Copenhagen Trial Unit, Centre for Clinical Intervention Research, Department 7812, Copenhagen, Denmark

  3. 3

    Rigshospitalet, Paediatric Department 4072, Copenhagen, Denmark

  4. 4

    Herlev University Hospital, The Cochrane Anaesthesia Review Group, Rigshospitalet & Department of Anaesthesiology, Herlev, Denmark

*Arash Afshari, The Cochrane Anaesthesia Review Group & Copenhagen Trial Unit and Department of Paediatric and Obstetric Anaesthesia, Rigshospitalet, Blegdamsvej 9, Afsnit 3342, rum 52, Copenhagen, 2100, Denmark. arriba.a@gmail.com. afshari@rocketmail.com.

Publication History

  1. Publication Status: Edited (no change to conclusions)
  2. Published Online: 16 JUL 2008

SEARCH

 
Characteristics of included studies [ordered by study ID]
Albert 1992

MethodsTwo-group parallel randomized clinical trial
Generation of allocation sequence: adequate
Allocation concealment: adequate
Blinding: unclear
Follow up: adequate
ITT: no
Overall study quality: high risk of bias
No sample size calculation was reported


Participants32 patients in the ICU with plasma AT III levels less than 70%. All patients included in the study within 2 days after admission to the ICU.

AT III group: 16 patients, 8 men, median age: 63.5 years. Median APACHE II score on inclusion: 14, TISS (therapeutic intervention scoring system): 34. Reason for admission: surgery 7 patients, trauma 4 patients, infection 3 patients, and other causes: 2 patients.

Control group: 16 patients, 9 men, median age: 66 years. Median APACHE II score on inclusion: 15, TISS (therapeutic intervention scoring system): 39. Reason for admission: surgery 5 patients, trauma 4 patients, infection 5 patients, and other causes: 2 patients.

Exclusion criteria: under 18 years old, liver disease, human immunodeficiency virus (HIV) infection, severe brain damage, patients admitted for post-operative pain relief only, anorexia nervosa, and pregnancy.


InterventionsAT III: concentrate by intravenous infusion for 20-30 minutes twice daily as long as AT III < 90%, in a dose intended to raise the level above 100% until the patients were discharged from the ICU. First dose: 2000-4000 IU, total amount: 3500-17000 IU.

Control: no placebo.

Unfractionated heparin 2500 or 5000 IU x 2 or x 3 daily was given subcutaneously to all patients for thromboprophylaxis. If continuous arteriovenous haemofiltration for acute anuria, then 20000-40000 IU Heparin per 24 hours was administered as continuous infusion. Blood samples from each patient were taken daily for 7 days or until discharge. Standard intensive care treatments were administered to all patients in both groups.


OutcomesPrimary: mortality.

Secondary: days in ICU, bleeding and transfusion, AT III, CRP, tissue plasminogen activator inhibitor (PAI-1), partial thromboplastin time, prothrombin complex (prothrombin time), fibrinogen degradation products (FDP), lactoferrin/serum, lactoferrin/plasma, lysozyme, side effects, APACHE II score and TISS (during the first 24 hours).


NotesCountry: Sweden
Letter sent to authors in November 2005. Reply received in December 2005.
2 patients randomized but not evaluated: 1 transferred to another hospital, the other had received AT III although randomized to the non-AT III group. There were 3 cases of cross over: these patients received AT III despite being allocated to the control group. These patients were evaluated up to the day when they received AT III. No side effects were reported.

Authors' conclusions: AT therapy appears to reduce hypercoagulation, fibrinolysis and the inflammatory response in seriously ill patients with acquired AT III deficiency.


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Low riskYes

Allocation concealment (selection bias)Low riskA - Adequate

Baudo 1992

MethodsTwo-group randomized clinical trial
Generation of allocation sequence: unclear.
Allocation concealment: unclear
Blinding: unclear
Follow up: adequate
ITT: yes
Overall study quality: high risk of bias
No sample size calculation was reported


Participants29 patients with post necrotic cirrhosis undergoing liver transplantation were randomized to receive AT III (13 patients) or to control group (no placebo, 16 patients) - before the induction of anaesthesia and during surgery.

AT III group: 13 men, 4 patients with cryptogenetic cirrhosis, 4 with post-necrotic cirrhosis, 5 with alcoholic cirrhosis, average age 44±9 years, duration of surgery (min): 625±68.
Control group: 16 men, 5 patients with cryptogenetic cirrhosis, 7 with post necrotic cirrhosis, 4 with alcoholic cirrhosis, average age 42±9 years, duration of surgery (min): 642±113.


InterventionsAT III: administered by bolus infusion before the induction of anaesthesia in order to obtain a pre-operative activity of 100%. Thereafter 1,000 units per hour were given by continuous infusion until the end of surgery. No systemic heparinization.

Control: no placebo.


OutcomesMortality. Blood loss, transfusion requirements, activation of coagulation and fibrinolysis (thrombin-antithrombin complexes), activated partial thromboplastin time (aPTT), prothrombin time (PT), thrombin time (TT), reptilase time (RT), expressed as ratio (R), fibrinogen g/l), plasminogen (plg), alfa2-antiplasmin, AT III, platelet count, thrombin-antithrombin complexes (TAT), total fibrinogen, fibrinogen and fibrin degradation products.


NotesCountry: Italy
Letter sent to authors in January 2006, no reply received.

Authors' conclusion: AT III therapy did not support the working hypothesis; AT III replacement therapy did not modify total blood loss and total transfusion requirements.


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Unclear riskUnclear

Allocation concealment (selection bias)Unclear riskB - Unclear

Baudo 1998

MethodsTwo-group parallel randomized clinical trial
Generation of allocation sequence: adequate
Allocation concealment: adequate
Blinding: adequate
Follow up: adequate
Overall study quality: low risk of bias
ITT: no
Sample size calculation reported


ParticipantsMulticentre trial, 120 patients admitted to the ICU with an AT III concentration of < 70%.

AT III group: 60 patients, 42 were men, age distribution: 58.6±13.8 years, multiorgan failure score (MOFS) 5.6±2.5, simplified acute physiology score (SAPS): 15.6±4.4, 49% sepsis, 33% septic shock, 51% respiratory support, 53% haemodynamic support, haemorrhages: 5%, surgery 47%, surgery with sepsis 43%, AT III: 52.8±15.5%.

Control group: 60 patients, 41 were men, age distribution 62±12.2 years, MOFS: 4.8 ±2.3, SAPS:16.5±5.5, 51% sepsis, 23% septic shock, 48% respiratory support, 42% haemodynamic support, haemorrhages: 6%, surgery 45%, surgery with sepsis 41%, AT III: 52.9±14.5%.

Inclusion criteria: 1) respiratory and/or haemodynamic support, 2) sepsis: systemic response to infection manifested by 2 or more of a) tp > 38.5 ºC, b) heart rate > 90/min, c) respiratory rate > 20/min, d) leukocytosis > 15 x 109/l or leukopenia < 4 x 109/l, e) septic shock with sepsis-induced hypotension requiring vasoactive drugs for more than 24 h persisting despite adequate fluid replacement along with presence of hypoperfusion abnormalities that may include but are not limited to lactic acidosis, oliguria, or an acute alteration of mental status, f) postsurgical complications requiring respiratory and/or haemodynamic support, g) respiratory support with assisted or controlled ventilation > 24 h, h) haemodynamic support inotropic (dopamine/dobutamine = / >5 mg/kg /min) and/or vasoactive (epinephrine or norepinephrine), 3) 18-75 years, 4) AT III activity < 70%.

Exclusion criteria: 1) Multiple trauma, 2) liver cirrhosis/acute liver failure, 3) Cancer in a terminal phase, 4) Immunodeficiency, 5) leukaemia, 6) pregnancy, 7) Heparin therapy (except prophylaxis heparin, 5000 units b.i.d., subcut.)


InterventionsAT III: total dose 24000 units, fixed dose of 4000 AT III and 2000 IU every 12 hours, 5 days, infusion by pump.

Control: human albumin, 50 g/l, 4 g albumin bolus in 30 min, 1 bottle containing 2 g albumin every 12 h for 5 days, pump-driven.

Standard intensive care treatment, fresh frozen plasma (FFP) infused when bleeding and/or prothrombin time ratio > 2, platelet concentration (PC) infused when </= 50 (1 unit/10 kg weight).


OutcomesPrimary: mortality, survival for 30 days, multiorgan failure score (MOF) for 7 days, FFP and PC requirements.

Secondary: days in ICU, bleeding and transfusion, MOFS, septic shock survival, AT III plasma concentration.


NotesCountry: Italy
Letter sent to authors in December 2005, no reply received.
Balanced randomization within each center, post hoc analysis of septic shock subgroup. The distribution of patients with septic shock on haemodynamic support were unbalanced in the two groups at admission. Simplified acute physiology score (SAPS) and MOF score recorded for each patient at admission.

Four patients received therapy for less than 24 h; 1 patient in the placebo group was transferred to another hospital after the bolus infusion and was considered not evaluable; 3 patients with septic shock (2 AT III and 1 placebo) died within 24 h. A total of 119 were included in the analysis. No adverse reaction was recorded.

Authors' conclusion: The results showed a non-significant reduction in mortality (50% in AT group vs 54% in the placebo group). AT III only reduces 30-day mortality significantly in the subgroup of septic shock patients (70% in the AT group vs 87% in the placebo, P<0.04). Because at inclusion, septic shock and haemodynamic support were unbalanced between the placebo and the treated groups, a post hoc analysis was done and showed that AT decreased the risk of death with OR of 0.56. Probability of survival was 30% in AT group vs 17% in placebo ( P<.05).


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Low riskAdequate

Allocation concealment (selection bias)Low riskA - Adequate

Diaz-Cremades 1994

MethodsTwo-group parallel randomized clinical trial
Generation of allocation sequence: inadequate
Allocation concealment: inadequate
Blinding: inadequate
Follow up: unclear
ITT: no
Overall study quality: high risk of bias
No sample size calculation was reported


Participants195 critically ill patients admitted to ICU, without previous history of hepatic disease, 36 patients were chosen with AT III levels below 70% as inclusion criteria. None of the patients had manifest DIC.

AT III: 20 patients (10 with sepsis, 6 with trauma, 1 with sepsis + trauma, 3 with shock), 12 patients were men, mean age 46.25±11.79 years.

Control group: 16 patients (8 with sepsis, 4 with trauma, 2 with sepsis + trauma and 2 with shock), 10 patients were men, mean age 56.37±25.78 years.

Inclusion or exclusion criteria were not provided.


InterventionsAT III group: initial dose 60 U/kg of body weight, then 10 U/kg every 6 h, without dose adjustment depending on the AT III levels. The mean AT III dose received was 11.17±5.98 U. Treatment interrupted when the patients were discharged from ICU, or if they achieved AT III levels equal or superior to 90% during at least 48 h. Mean length of treatment: 5.8±4.2 days.

Control: human albumin 0.6%.

Standard intensive care treatment in both groups. All patients received 7500 IU of prophylactic subcutaneous calcium heparin every 12 h.


OutcomesPrimary: mortality.

Secondary: 1) platelet count, 2) side effects, 3) prothrombin time (PT), 4) activated thromboplastin time (APTT), 5) fibrinogen, 6) fibrinogen degradation products, 7) AT III, 8) protein C, 9) APACHE II score.


NotesCountry: Spain
Letter sent to authors in December 2005, no reply received.

No adverse effects reported. The only detectable alterations in the haemostasis were a mild thrombocytopenia and low levels of AT III and protein C. Manifest DIC was not reported in any patient.

Authors' conclusion: administration of AT III to critical patients with acquired low levels but without manifest DIC neither contributes to alterations in haemostasis nor the clinical evolution.


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)High riskInadequate

Allocation concealment (selection bias)High riskC - Inadequate

Eisele 1998

MethodsRandomized clinical trial
Generation of allocation sequence: unclear
Allocation concealment: unclear
Blinding: unclear
Follow up: yes, 30 days or until death
Overall study quality: high risk of bias
ITT: yes
No sample size calculation was reported


ParticipantsMulticentre, multinational trial of patients admitted to ICU.

AT III group: 20 patients, 11 were men, mean age: 57±16 y, weight (kg): 71±14, height (cm): 171±9. Underlying disease: peritonitis 2, mediastinitis 1, neoplasia (non-hematologic) 4, others: 12. Severity of sepsis at study entry (mean ± SD): APACHE II: 13.5 ± 5.1, MOF: 6.1±2.1, OFS:1.2±0.9, AT III activity: 45.7±14.4%. Coagulation failure: 3, acute hepatic dysfunction: 5, acute renal dysfunction: 8, acute respiratory dysfunction: 8, acute neurological dysfunction: 6.

Control group: 22 patients, 10 were men, mean age: 58±14 y, weight (kg): 73±15, height (cm): 169±8. Underlying disease: peritonitis 3, mediastinitis 1, adult respiratory distress syndrome 1, acute pancreatitis 3, neoplasia (non-haematologic) 3, other: 11. Severity of sepsis at study enter (mean ± SD): APACHE II: 15.5 ± 5.7, MOF: 6.7±2.0, OFS: 0.9±0.8, AT III activity: 49.0±19.1%. Coagulation failure: 2, acute hepatic dysfunction: 8, acute renal dysfunction: 8, acute respiratory dysfunction: 6, acute neurological dysfunction: 10.

Inclusion criteria: less than 80 and above 18 years old; fulfilled the following criteria within 6 h window prior to initiation of treatment: 1) clinical evidence of sepsis with a suspected source of infection; 2) temperature rectally > 38.5 ºC or < 35.5 ºC; 3) leukocyte count >10000/mm3 or <3500. Furthermore 3 of the following had to be fulfilled: a) heart rate >100/min; b) breath rate >24/min or mechanical ventilation because of septic indication; c) hypotension, sys BP <90 mm Hg when no vasoactive agent or fluid replacement; d) platelet count <100000/mm3; e) lactate >1.6 mmol/l; f) urine output <20ml/h.

Exclusion criteria: a) heparin (except sc low-dose), b) malignancies (incurable with metastases); c) haematologic neoplasia (except complete remission); d) chronic treatment with high dose immunosuppressive drugs; e) high dose NSAID within previous 2 days; f) known bleeding disorder; g) ongoing massive surgical bleeding, h) multiple organ failure, at least 4 organs involved existing for over 24 h; i) acute myocardial Infarction; j) chronic compensated organ dysfunction (chronic liver disease, dialysis, NYHA 3-4); k) complicated pre-existing diabetes mellitus; l) severe obstructive pulmonary disease; m) burn; n) AIDS; o) transplant, p) severe cranial trauma and/or pathologic changes on CT and GCS < 6; q) pregnancy; r) treatment with plasma expanders (Haes); s) prior enrolment in the trial or another clinical trial within the last 30 days.


InterventionsExperimental: AT III iv loading dose: 3000 units AT III, maintenance 1500 IU every 12 hours, 5 days treatment, cumulative dose: 18000 IU.

Control: placebo equivalent amounts, unknown agent.

Standard intensive care treatment in both groups.


OutcomesPrimary: 30-day all-cause mortality.

Secondary: survival days, days in ICU, safety (bleeding and transfusion), MOFS, APACHE II, OSFS, AT III activity, AT III plasma concentration, resolution of acute pre-existing organ dysfunction (neurological, respiratory, renal, hepatic and coagulation) and incidence of organ dysfunction.


NotesCountry: Germany, Belgium, the Netherlands
A letter sent to authors in December 2005, no answer received.

Acute physiology and chronic health evaluation (APACHE II), multiple organ failure (MOF) and organ system failure (OSF) scores were calculated on days 1, 2, 3, 7, 10, 20 and 30.
No adverse events reported.

Authors' conclusion: AT III administration was safe and well tolerated. 39% reduction in 30 day all-cause mortality. Shorter stay in ICU. Better performance in overall severity of illness and organ failure scores (APACHE II, MOFS, OSF score) noticeable soon after initiation of treatment. Trend towards better resolution of pre-existing organ failures and lower incidence of new organ failures. From day 4: AT III within normal range in treatment group. Survival benefit appeared to occur late in the septic process (second week after initiation of treatment, most pronounced during the third week). No advantage in regard to coagulation.


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Unclear riskUnclear

Allocation concealment (selection bias)Unclear riskB - Unclear

Fourrier 1993

MethodsTwo-group parallel randomized clinical trial
Generation of allocation sequence: adequate
Allocation concealment: adequate
Blinding: adequate
Follow up: yes, 28 days or until death
Overall study quality: low risk of bias
ITT: yes
Sample size was calculated to detect a 50% reduction in mortality in ICU


ParticipantsCritically ill patients with septic shock and DIC and MOF with at least 1 organ failure added to circulatory and hematologic failures.

AT III group: 17 patients, 10 were men, mean age (SD): 52±22 y. Underlying disease: alcoholism 5; chronic cardiac failure 8; COPD 2. No underlying disease 3; recent surgery 6. SAPS: 20±7. Organ failures: 2.8±1.0, renal failure: 57%, respiratory failure: 93%.

Control group: 18 patients, 10 were men, mean age (SD): 52±18 y. Underlying disease: alcoholism 3; chronic cardiac failure 9; COPD 3. No underlying disease 4; recent surgery 6. SAPS: 19±6. Organ failures: 2.8±0.9, renal failure: 56%, respiratory failure: 89%.

Inclusion criteria: A) mandatory criteria for septic shock: 1) bacteremia or evidence of septic focus, 2) fever >38.5 ºC, 3) leukocyte count >15x109/L or <4x109; 4) hypotension, sys BP <90 mmHg despite adequate vascular filling; 5) requirement of vasoactive agent (dopamine >15 µg/kg/min, dobutamine, epinephrine or norepinephrine) during more than 12 h.
B) Mandatory DIC-criteria: 1) platelet counts <100x109/L or a decrease of >100x109 over the last 24 hours; 2) decrease in factor V level <70%; 3) presence of soluble complexes of fibrin monomers.

Exclusion criteria: a) younger than 16 years old; b) suspected pregnancy; c) sepsis or DIC within 8 days after delivery; d) chronically immunocompromised; e) bone marrow aplasia or acute leukaemia; f) chronic liver failure with cirrhosis, recent hepatic surgery or transplant; g) multiple trauma or uncontrolled haemorrhage that required massive blood transfusion or undergone plasma exchange (< 5 days).


InterventionsExperimental: 90-120 IU/kg, loading dose over 3 h (3 ml/kg) + 3 ml/kg over 21 h, then 90-120 IU/kg/day for 3 days by continuous infusion. The aim: supranormal AT III levels. Average of 6000 IU/day.

Control: human albumin 0.6%, equivalent amount in ml.

Standard intensive care treatment to all patients. FFP, PC, and fibrinogen concentrates restricted to haemorrhages and severe decreases in prothrombin time to <30%, platelet count to <50x109/L (1/10 kg of body weight), and fibrinogen levels to <1 g/L. LMWH 0.5 mg/kg of body weight when used in lines and filters of haemodialysis or haemofiltration.


OutcomesPrimary: 28-day mortality.

Secondary: recovery of DIC, side effects, length of stay, transfusion and vascular requirements, bleeding, numbers and types of organ failure, organ function scores, protein C+S activity, fibrinogen and AT III activity, fibrin degradation products, factor V and VII, cardiac index, systemic vascular resistances, blood lactate, total bilirubin, blood creatinine, omega score, surgery during treatment.


NotesCountry: France
Letter sent to authors in November 2005, authors replied in December 2005.

Heparin therapy was not used as control to avoid haemorrhagic hazards and discrepancies related to heparin-induced AT III consumption. Since FFP contains significant amount of AT III, infusion was restricted. No side effects reported.

Authors' conclusion: mortality reduced by 44% but it did not reach statistical significance. AT III levels were rapidly corrected and remained normal until day 10. Circulating protein C and S levels were not modified by AT III supplementation. Duration of DIC was reduced significantly (64% after day 2; 71% at the end of treatment vs placebo 11% (P<0.01) and 33 (P< 0.05)).


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Low riskAdequate

Allocation concealment (selection bias)Low riskA - Adequate

Fulia 2003

MethodsTwo-group parallel randomized clinical trial
Generation of allocation sequence: adequate
Allocation concealment: adequate
Blinding: adequate
Follow up: yes
Overall study quality: low risk of bias
ITT: yes
No sample size calculation was reported


Participants60 infants born before 30 weeks of gestation randomized to a loading dose of AT III or placebo.

AT III: 30 infants, 16 were male, gestational age (weeks): 28.5±1.5, birth weight (g): 1060±218, vaginal delivery: 12, elective caesarean: 18, antenatal steroids: 16, APGAR score at 1 min: 5.83±0.75, APGAR score at 5 min: 7±0.64, cord ABG pH: 7.28±0.07.

Control: 30 infants, 17 were male, gestational age (weeks): 28.5±1.3, birth weight (g): 1054±233, vaginal delivery: 11, elective caesarean: 19, antenatal steroids: 17, APGAR score at 1 min: 5.63±0.82, APGAR score at 5 min: 6.88±0.76, umbilical cord arterial blood gas pH: 7.31±0.08.

Inclusion criteria: gestational age <30 weeks, postnatal age <12 h, AT III activity <40% (normal value 20-72%).

Exclusion criteria: sepsis, congenital malformations, cerebral haemorrhage, bleeding disorders, thrombocytopenia (platelet count of 50 x 109 or less).


InterventionsAT III group: loading dose of 2 ml/kg equivalent to 100 U/kg of AT III followed by 1 ml/kg equivalent to 50 U/kg every 8 h for 48 h.

Control: 5% glucose in equivalent amounts (ml).

Arterial catheters were perfused with 1 ml/h of 5% glucose containing I U/ml of unfractionated heparin. No other anticoagulants were administered. All newborns received the same exogenous surfactant.


OutcomesPrimary: death and intraventricular haemorrhage (grade 0-IV).

Secondary: partial thromboplastin time, quick prothrombin time, platelet count, surfactant need, pneumothorax, pulmonary haemorrhage, patent ductus arteriousus, inotropes, bronchoplumonary dysplasia, level of AT III, degree of IVH measured by ultrasonography.


NotesCountry: Italy
Letter sent to authors in January 2006, no answer received.

Authors' conclusion: administration of AT III during the first 2 days of life does not decrease incidence of intraventricular haemorrhage.


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Low riskAdequate

Allocation concealment (selection bias)Low riskA - Adequate

Grenander 2001

MethodsTwo-group parallel prospective randomized trial
Generation of allocation sequence: unclear
Allocation concealment: adequate
Blinding: no
Follow up: adequate, evaluation at discharge and at 3 months post injury
Overall study quality: high risk of bias
ITT: no
No sample size calculation was provided


Participants28 patients with purely traumatic brain injury and no other major trauma

AT III group: 13 patients, 10 were men, mean age: 48 years. Mean APACHE II score: 14.2, mean GCS: 7.7, median: 7.0; 50% had positive blood alcohol detection.

Control group: 15 patients, 10 were men, mean age: 44 years. Mean APACHE II score: 14.8, mean GCS: 7.5, median: 7.0; 50% had positive blood alcohol detection.

Inclusion criteria: 1) 14-70 years old, 2) brain injury was CT-verified and assessed by Glasgow coma scale (GCS), 3) only GCS between 4-12 included, 4) inclusion within 12 h from the time of injury.

Exclusion criteria: 1) ongoing dialysis-dependent renal disorder, 2) bleeding or coagulation disorder such as haemophilia or von Willebrand's disease, 3) anticoagulant therapy.


InterventionsAT III: 60 IU/kg body weight initially. 8 and 16 hours later an additional 20 IU/kg body weight was given - a total of 100 IU/kg body weight during 24 h. The doses were adjusted to the nearest 500 IU. Duration: 24 h. Average total dose of AT given was 8269±1562 IU.

Control: no placebo.

Both groups received additional treatment according to local standards for brain injury including surgery, ventricular drain, or intraparenchymatous intracranial pressure (ICP) device, moderate hyperventilation (pCO2 4-4.5 kPa), sedation, drainage of cerebrospinal fluid, low-dose barbiturates, mannitol infusion, inotropic drugs, clonidine and dihydroergotamine. Treatment goals were ICP<20 mm Hg and CPP>60 mm Hg.
Low molecular weight heparin 2500 IU sc only to a random group of patients in few events (1 patient in AT group and 3 patients in control group). 3 patients in each group received fresh frozen plasma for volume replacement.


OutcomesPrimary outcome: time of coagulopathy in patients with brain contusion and with or without intracranial haemorrhage, coagulation parameters improvement: soluble fibrin (SF), AT III-activity, D-dimer, thrombin-antithrombin complex (TAT), platelet count, prothrombin complex (PK) and fibrinogen, mortality at 3 months.

Secondary: duration of mechanical ventilation, APACHE II score (slightly modified in order to estimate the severity of illness), improved outcome assessed by GCS (once daily and at 3 months), brain injury progress monitored by CT, bleeding events, days with mechanical ventilation, need for intensive care, and evaluation of severe disability.


NotesCountry: Sweden
Letter sent to authors in December 2005, no answer received.

Two withdrawals (late presentation of exclusion criteria and a revoked consent), one late exclusion resulting from the revelation of improper dosage of antithrombin because of incorrect information regarding body weight. One case of cross over. One patient in the control group was given AT III by the attending physician as part of the treatment for a septic reaction. Partially missing data on this patient.

Authors' conclusion: AT III administration to patients with severe traumatic brain injury resulted in a marginal reduction of hypercoagulation. No obvious influence by AT III on brain injury progress, CT, or on outcome or time needed for intensive care could be detected.


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Unclear riskUnclear

Allocation concealment (selection bias)Low riskA - Adequate

Haire 1998

MethodsTwo-group parallel randomized clinical trial
Generation of allocation sequence: adequate
Allocation concealment: adequate
Blinding: adequate
Follow up: adequate, until hospital discharge or death
Overall study quality: low risk of bias
ITT: yes
Sample size calculation reported


Participants49 patients undergoing haematopoietic stem cell transplantation (HSCT).

AT III group: 24 patients, 10 were men. Mean age (SD): 46.1±10.3 y. Underlying disease: acute myeloid leukaemia 5, breast carcinoma 8, chronic myeloid leukaemia 1, multiple myeloma 3, lymphoma 7.

Control group: 25 patients, 11 were men. Mean age (SD): 46.0±10.5 y. Underlying disease: acute lymphocytic leukaemia 1, acute myeloid leukaemia 4, breast carcinoma 4, chronic lymphocytic leukaemia 2, chronic myeloid leukaemia 4, Hodgkin's disease 1, multiple myeloma 3, lymphoma 6.

Inclusion criteria: malignant disease admitted for haematopoietic stem cell transplantation (HSCT), not committed to other confounding clinical trials, English their first language (to allow reliable use of the mini mental status examination).


InterventionsAT III: 70 U/kg within 24 h of organ dysfunction detection, followed by 50 U/kg 8,16, 48 and 72 hours later. Total dose of AT III: 270 U/kg. Estimated to provide an increment of about 250% activity above pretreatment levels after the 3rd dose. Mean total dose: 20,520 U/patient based on the mean weight of the randomized patients (76 kg).

Control: 5% human albumin, volume equal to the calculated volume of AT III concentrate.


OutcomesPrimary: mortality, severity-of-illness score, length of hospital stay.

Secondary: daily screening for pulmonary, hepatic or CNS dysfunction, MODS (multiorgan dysfunction syndrome). Correlation between AT III, number of organ failure and MODS. Number of patients progressing from single through multiple organ dysfunction.


NotesCountry: USA
Letter sent to authors in December 2005, no answer received.
One patient was randomized and received the first dose of the study drug based on an erroneous laboratory report. This patient was not included in the final analysis.
Organ dysfunction defined as: 1 CNS - a drop in the score of standardized mini mental status examination of at least 4 points from the pre-chemotherapy score; 2) pulmonary - finger oximetry reading of SpO2 <90% on two occasions on the same day separated by at least 2 hours; 3) pepatic - a combination of bilirubin >2.0 mg%, a weight gain of >5% over pre-chemotherapy weight, and abdominal pain of possible hepatic origin.
Patients with single-organ dysfunction with a concomitant AT III activity of <84% of normal were defined as having MODS.

Authors' conclusion: the AT III group had a lower severity-of-illness score, shorter duration of hospitalization and lower hospital charges. AT III concentrate was associated with improved morbidity of MODS in patients undergoing HSCT when given early in the evolution of the syndrome.


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Low riskAdequate

Allocation concealment (selection bias)Low riskA - Adequate

Harper 1991

MethodsTwo-group parallel randomized controlled trial
Generation of allocation sequence: unclear
Allocation concealment: inadequate
Blinding: inadequate
Follow up: inadequate
Overall study quality: high risk of bias
ITT: yes
No sample size calculation was provided


ParticipantsPatients with AT III activity of <70% admitted to the ICU. Patients with an initial AT activity above 70% monitored daily and were randomized into the study if their AT activity at any time during their stay fell below 70%.

AT III group: 44 patients, median age: 49 (range 19-78) y. Admission AT III activity: 51% (range 12-68); APACHE II score median: 16 (range 6-33).

Control group: 49 patients, median age: 49 (range 16-77) y. Admission AT III activity: 50% (range 12-64); APACHE II score median: 14 (range 4-26).

Exclusion criteria: acute or chronic renal failure prior to admission to the ICU. If readmitted to ICU, the patients were allocated to the same randomization as on their previous admission. Patients were excluded if they stayed less than 4 days on ICU.
10 liver transplant recipients and 15 patients with septicaemia, major trauma, or following major surgery in each group.


InterventionsTreatment: AT III, aiming at 120%. Loading dose was calculated from the initial AT III activity and the patient's weight ((120-AT III activity (%)) x weight (kg)). Maintenance doses were given iv over 10 min. Treatment was continued until the patient was discharged from the ICU, twice daily.

Control: no placebo.

Standard intensive care treatment to all patients.


OutcomesPrimary: mortality, coagulation parameters (AT III activity, AT-antigen concentration, fibrinogen concentration, platelet count). Renal function impaired if creatinine clearance <20 ml/min at any time during their stay on the ICU.

Secondary: days in ICU, side effects, APACHE II.


NotesCountry: UK
Letter sent to authors in November 2005, no answer received. No reported side-effects.

81 patients had an AT III activity below 70% on admission to the ICU and in a further 12 patients the plasma AT III fell below 70% whilst on the ICU. Thus a total of 93 patients were randomized: 44 to the treatment arm and 49 in the no-treatment arm. 9 patients in the AT III arm did not receive AT III replacement; 3 died before treatment was commenced; and 6 were transferred from the ICU within a few hours of randomization. A total of 35 patients received AT III replacement. Of these 25 remained on the ICU for at least 4 days. 32 patients in the control arm remained on the ICU for at least 4 days. 25 of these were selected by diagnosis to act as controls. The authors provided mortality data on all patients randomized, and we have chosen to include these in our analysis.

Authors' conclusion: AT III neither reduced mortality nor shortened the intensive care stay. AT III did not appear to reduce the incidence of impaired renal function in sepsis, trauma and post-operative patients. The study does not demonstrate a clear role for the use of AT III supplementation in intensive care.


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Unclear riskUnclear

Allocation concealment (selection bias)High riskC - Inadequate

Inthorn 1997

MethodsTwo-group parallel randomized clinical trial
Generation of allocation sequence: unclear
Allocation concealment: unclear
Blinding: unclear
Follow up: yes
Overall study quality: high risk of bias
ITT: yes
Sample size calculation reported


Participants40 surgical patients with severe sepsis, post-traumatic or post-operative admitted to surgical ICU.

AT III group: 20 patients. Demographic characteristics only provided for the 14 patients included: 13/14 were men. Source of sepsis: peritonitis (10), pneumonia (2), catheter-related infection (1), late abscess (1). Age: 62.3± 3.8 y, MOF score: 4.1± 0.5.

Control group: 20 patients. Demographic characteristics only provided for the 15 patients included: 11/15 were men. Source of sepsis: peritonitis (12), pneumonia (2), Mediastinitis (1). Age: 61.5± 3.6 y, MOF score: 3.9± 0.4.

Inclusion criteria: a) one or more blood cultures positive for gram positive or negative bacteria or evidence of septic focus; b) temperature above 38.5 ºC, c) leukocyte count >15 g/L or <5 or >20% of immature (band) forms; d) thrombocyte count <100 g/L or a decrease of >20% during the last 24 hours; e) at least one organ dysfunction.

Exclusion criteria: 1) <18 years old; 2) received AT within 21 days prior to the study.


InterventionsAT III: continuous AT III infusion over 14 days in order to obtain plasma AT III > 120%. Average of 6000 IU AT III was administered on 1st day and 4000 IU on the subsequent days.

Control: no placebo treatment.

All patients received standard relevant intensive care treatment that was comparable between groups and continuous low-dose heparinization (4 IU/kg bodyweight/h). FFP was only given to patients with severe haemorrhage and markedly impaired plasmatic coagulation (contains AT III); prothrombin complex given only during excessive clinical bleeding; platelets given only when <50 g/L; packed red blood given when <10 g/dL.


OutcomesPrimary: 14 and 90-day mortality; mortality between days 7 to14, hospital discharge.

Secondary: transfusions, MOF Score, frequency of surgical interventions, frequency of DIC, AT III plasma activity, severity of organ dysfunction, effect of AT III on clotting system and MOF, haemodynamic and respiratory profile.


NotesCountry: Germany
Letter sent to authors in November 2005, no answer received.
The results of this study were also published in two additional articles (Inthorn 1998 and Hoffmann 2004 under the excluded articles section). Comparable patients in both groups and no significant baseline differences in measured outcomes. Peritonitis most frequent source of sepsis (70%).

Due to the presumed mechanisms of AT III action, only long-term supplementation was assumed to affect parameters of organ function. Thus, 11 patients were prospectively excluded from the final evaluation since they died during the 14-day period. However their data were included in an intention-to-treat analysis. No side effects were reported.

The patients in both groups were comparable with no significant baseline differences in measured outcomes. Peritonitis was the most frequent source of sepsis (70%).

Authors' conclusion: long-term supplementation with AT III may improve lung function and prevent development of septic liver and kidney failure in patients with severe sepsis.


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Unclear riskUnclear

Allocation concealment (selection bias)Unclear riskB - Unclear

Kobayashi 2003

MethodsTwo-group parallel randomized clinical trial
Generation of allocation sequence: adequate
Allocation concealment: adequate
Blinding: adequate
Follow up: adequate
Overall study quality: low risk of bias
ITT: yes
Sample size calculation reported


Participants29 severe pre-eclamptic patients (24 to 36 weeks of gestation, gestosis index (GI = or >6). Severe pre-eclampsia was defined by the presence of hypertension plus proteinuria.

AT III: 14 women; age: 30±6 y; pregnancy weight (kg): 51±7 SD; systolic BP (mm Hg): 163 ± 21, diastolic BP (mmHg): 104±15; gestosis index: 7.2±1.6; gestational age (weeks): 32.3±3.1; biophysical profile score: 7.0±2.0; primigravida (%): 71.4.

Control: 15 women; age: 30±4 y; pregnancy weight (kg): 53±9; systolic BP (mm Hg): 165 ± 12, diastolic BP (mmHg): 100±14; gestosis index: 6.6±1.8; gestational age (weeks): 29.8±3.7; biophysical profile score: 7.4±1.9; primigravida (%): 60.0.

Inclusion criteria: systolic BP = or >160 mm Hg and/or diastolic BP > or = 110 mm Hg on 2 occasions 6 hours or more apart; and/or proteinuria > or = 2 g/L in a 24-hour urine collection; and a GI = or >6 points on 2 occasions at least 6 hours apart at bed rest.

Gestosis index score was calculated as: A) edema after bed rest: 0 (none), 1 (tibia), 2 (generalized); B) proteinuria (g protein/L) 0 (<0.5), 1 (> or = 0.5-1.99), 2 ( > or = 2.0-4.99), 3 (> or = 5.0); C) systolic BP (mm Hg) 0 (<140), 1 (140-159), 2 (160-179), 3 (>180); D) diastolic (BP mm Hg) 0 (<90), 1 (90-99), 2 (100-109), 3 (110).

Exclusion criteria: patients with chronic hypertension, renal disease, diabetics mellitus, systemic lupus erythematosus, multiple pregnancies, patients with AT deficiency and other severe medical conditions.


InterventionsAT III: 1500 U/day once daily for 7 consecutive days.

Control group: unknown placebo.

5000 U/day of unfractionated heparin was given simultaneously in both groups. Heparin was infused by 24-hour intravenous drip with saline.
Concomitant therapy with other anticoagulants, antiplatelet agents, and blood preparations except albumin was not permitted during the study. Only hydralazine hydrochloride at a 30 mg/day was allowed in both groups.


OutcomesNeonatal and fetal outcomes: mortality (neonatal and fetal), bleeding disorder, weeks of gestation at delivery, birth weight (g), APGAR score at 1 minute, neonatal distress (%), response on the biophysical profile score.

Maternal outcomes: gestosis index improvement, improvement of coagulation index, adverse events, blood loss.


NotesCountry: Japan
Letter sent to authors in December 2005, reply received in January 2006.
No adverse event related to AT III was reported.

The criteria used to initiate a preterm delivery were based on sustained severe hypertension (> or =180/110 mm Hg), thrombocytopenia (< or = 100 x (109/L), markedly worsened liver or renal function (serum aspartate aminotransferase or serum alanine amino transferase > or = 100 IU/L, serum creatinine > or = 15 mg/L, blood urea nitrogen > or = 200 mg/L), severe maternal symptoms (eclampsia, HELLP syndrome, placental abruption, pulmonary edema, etc) or impaired fetal status evidenced by intrauterine growth retardation (< or = -2 standard deviations) and oligohydramnios (amniotic pocket > or = 1 cm), or persistently severe abnormal fetal heart rates or worsened biophysical profile test results (based on fetal breathing movements, fetal tone, reactive fetal heart rate, gross body movements, qualitative amniotic fluid volume).

Authors' conclusion: Given that AT plus heparin was significantly better than treatment with heparin alone for improvement of both gestosis Index and biophysical profile of infant, it is suggested that therapy with AT alone might be effective enough for severe pre-eclampsia.


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Low riskAdequate

Allocation concealment (selection bias)Low riskA - Adequate

Langely 1993

MethodsTwo-group parallel randomized clinical trial
Generation of allocation sequence: unclear
Allocation concealment: unclear
Blinding: unclear
Follow up: adequate
Overall study quality: high risk of bias
ITT: yes
Sample size calculation was not reported


Participants25 patients in grade III or IV coma were selected on the basis of evidence of sepsis, intravascular coagulation, and a high risk of developing multiorgan failure.

AT III: 13 patients, 7 were men, 12 with severe hepatic dysfunction on entry of which 9 also had further complications of fulminant hepatic failure (FHF). Underlying etiology: paracetamol overdose (9), hepatotoxicity to phenytoin (1), hepatitis B (1), non-A non-B hepatitis (2). Time from the onset of illness to entry into the trial: median 3 (2-28) hours.

Control: 12 patients, 4 were men, all with severe hepatic dysfunction on entry of which 6 also had further complications of FHF. Underlying etiology: paracetamol overdose (8), (1), hepatitis B (2), non-A non-B hepatitis (1), halothane hepatitis (1). Time from the onset of illness to entry into the trial: median 3 (2-37) hours.

No statistically significant differences between the two groups on admission with respect to age, sex, prothrombin ratio, and initial AT III level.

Inclusion criteria: 1) severe hepatic dysfunction, prothrombin ratio expressed as an international normalized ratio (INR) >6.7 for paracetamol overdose or INR >3.3 for other etiologies; 2) evidence of sepsis with temperature higher than 39 ºC, white cell count >15 x 109/L, or microbiologically proven sepsis; 3) presence of disseminated intravascular coagulation with spontaneous bleeding and a platelet count of less than 50 x 109/L; 4) multiorgan failure with systemic blood pressure <80 mm Hg, requiring inotropic support to maintain blood pressure and blood pH equal to or less than 7.3.

Exclusion criteria: not provided.


InterventionsAT III group: 13 patients received 3000 units of AT III on entry into the trial followed by a further 1000 units every 6 hours unless AT III levels were within the normal range. Total amount of AT III given to the patients during the course of their illness ranged from 3000 units to 23,000 U (mean 7000± 5000 U).

Control: no placebo.

All patients received standard supportive therapy for FHF witch included haemodynamic monitoring, haemodialysis or haemofiltration for renal failure, and continuous intracranial monitoring with treatment for raised intracranial pressure where necessary. Mechanical ventilation was used to treat respiratory dysfunction or pulmonary complications.


OutcomesPrimary: mortality, plasma levels of AT III.

Secondary: renal failure, cerebral edema, incidence of infection, liver transplantation, coma grade, effect of AT III on bleeding and coagulation parameters, transfusion requirements.


NotesCountry: United Kingdom
Letter sent to authors in December 2005, no reply received.

Authors conclusion: AT III activity increased from 0.26±0.04 SE U/ml to 0.82±0.07 U/ml at 3 hours post-infusion and remained greater than 0.80 U/ml throughout the study without any apparent increase in the frequency of bleeding. Survival was not improved and markers of intravascular coagulation remained similar between the groups.


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Unclear riskUnclear

Allocation concealment (selection bias)Unclear riskB - Unclear

Maki 2000

MethodsTwo-group parallel randomized clinical trial
Generation of allocation sequence: adequate
Allocation concealment: adequate
Blinding: adequate
Follow up: adequate, 14 day period including the time of drug administration, and until delivery to assess the birth and the status of the neonate
Overall study quality: low risk of bias
ITT: yes
No sample size calculation was reported


ParticipantsSevere pre-eclamptic patients (24 to 35 weeks of gestation).

AT III: 74 women, age: 30±5 y; pregnancy weight (kg): 66±12; systolic BP (mm Hg): 170±16, diastolic BP (mmHg): 103±12; gestosis index: 7.7±1.4; gestational age (weeks): 31.8±3.2; biophysical profile score: 8.6±2.0; primigravida (%): 34.8.

Control: 72 women, age: 30±5 y; pregnancy weight (kg): 65±12; systolic BP (mm Hg): 169±17, diastolic BP (mmHg): 105±12; gestosis index: 7.7±1.6; gestational age (weeks): 31.7±2.7; biophysical profile score: 8.7±1.7; primigravida (%): 37.3.

Inclusion criteria: systolic blood pressure (BP) >= 160 mm Hg and/or diastolic BP >= 110 mm Hg on 2 occasions 6 h or more apart; and/or proteinuria =>2 g/L of protein in a 24-h urine collection; gestosis Index, GI =/>6 points on 2 occasions at least 6 h apart despite bedrest. GI is calculated based on edema, proteinuria, systolic BP and diastolic BP.

Exclusion criteria: chronic hypertension, multiple pregnancy, renal disease, diabetics mellitus, systemic lupus erythematosus, other severe medical conditions and patients with AT deficiency.

Similar baseline characteristics, proportion of severe early onset pre-eclampsia was equally distributed within the study groups.
For definitions of gestosis index and bio-profile scoring technique, please see Kobayashi 2003 characteristics.


InterventionsTreatment: 3000 IU intravenous AT III once daily for 7 consecutive days.

Control: 582 mg albumin intravenously, once daily for 7 consecutive days as placebo.

Concomitant therapy with other anticoagulants, antiplatelet agents and blood preparations except albumin was not permitted during the study. Other relevant treatment such as antihypertensive agents, magnesium sulfate allowed in both groups. No heparin was used.


OutcomesPrimary endpoints: improvement of gestosis index, improvement fetal findings (biophysical profile score, and the estimated fetal weight gain), duration of pregnancy, birth weight, gestational age at delivery, improvement of coagulation parameters (TAT-complexes, plasmin-plasmin inhibitor complexes (PPIC), PC, D-dimer, AT activity, AT antigen), stillbirth or neonatal death ascribed to pre-eclampsia, maternal death from any cause, or death of neonate at any time attributed to pre-eclampsia or associated with intra-uterine growth retardation.

Secondary: adverse events related to AT.
Stillbirth included all intrauterine deaths at or after 24 weeks, and neonatal death included all deaths after birth up to the age of 28 days.


NotesCountry: Japan
Letter sent to authors in December 2005, reply received in January 2006.
Intervention had to be stopped in 23 patients in the AT III group and 29 patients in the control group prematurely due to worsening of maternal and fetal findings.

Authors' conclusion: AT therapy for pre-eclampsia is effective and safe, leading to an improved perinatal and maternal outcome. Significant improvement in GI (P=0.020) and estimated fetal weight gain (P=0.029) in AT group. D-dimer increased significantly in the placebo group (P=0.026) but did not change in the AT group. Gestation was significantly prolonged (P=0.007) and the number of low-birth weight infants was significantly smaller (P=0.011) in the AT-group. No adverse events related to AT. No significant difference was noted in Apgar score, neonatal distress, endotracheal intubation, fetal death, neonatal death, bleeding disorder, or maternal outcome.


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Low riskAdequate

Allocation concealment (selection bias)Low riskA - Adequate

Mitchell 2003

MethodsTwo-group parallel randomized clinical trial
Generation of allocation sequence: adequate
Allocation concealment: adequate
Blinding: no blinding
Follow up: inadequate.
Overall study quality: high risk of bias
ITT: no
Sample size calculation reported


ParticipantsPaediatric acute lymphoblastic leukaemia (ALL) patients.

AT III: 25 patients, 10 female, median age 3.8 (1.6-17.2) years, weight 18.3 (9.7-76.7) kg.

Control: 60 patients, 23 female, median age 5.9 (1.9-16.7) years, weight 20.6 (10.8-83.4) kg.

Inclusion criteria: age >6 months and <18 years, newly diagnosed with ALL at the beginning of the induction of chemotherapy, a functioning central venous line placed within 2 weeks of initiating induction chemotherapy and obtaining informed consent.

Exclusion criteria: previous treatment with L-asparaginase (ASP), a known hypersensitivity to any of the ingredients in antithrombin concentrate, medical conditions that could have interfered with participation or assessment of the study drug, received other investigational drugs within 30 days of enrolment, or required treatment with therapeutic anticoagulation.

All patients received unfractionated heparin for prophylaxis of central venous line-blockage either by continuous infusion (1-3 units/mL) or intermittent flushes (50-100 units/mL up to 4 times per day). The decision to use additional anticoagulants was left to the patient's physician.


InterventionsAT III: infusions once weekly for 4 weeks (days 1, 8, 15, 22) to increase plasma concentrations of AT to approximately 3.0 units/mL but no more than 4.0 units/mL.

Control: no placebo.


OutcomesPrimary: prevalence of thrombotic events.

Secondary: bleeding events, plasma markers, efficacy and safety.


NotesLetter sent to authors in December 2005, no answer received.
Country: Canada and USA
No data were provided on mortality. Based on the detailed follow up, we concluded that there was no mortality in both groups.

The study was not powered to prove efficacy or safety of AT III supplementation but rather look for trends. Outcome measures were assessed blinded.

24 patients excluded from analysis (12 in each group). The reasons for exclusion in the AT III group were: 3 patients due to withdrawal of consent, 2 patients categorised as noncompliant due to an under-dosing with AT concentrate, 7 patients due to the absence of an exit venogram. In the control group: 3 patients due to withdrawal of consent, 1 patient had an adverse event, and 8 patients were excluded due to the absence of an exit venogram.

Authors' conclusion: treatment with AT III shows a trend to efficacy and safety. In contrast, there was no difference in surrogate markers for thrombosis.


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Low riskAdequate

Allocation concealment (selection bias)Low riskA - Adequate

Schmidt 1998

MethodsTwo-group parallel randomized clinical trial
Generation of allocation sequence: adequate
Allocation concealment: adequate
Blinding: adequate
Follow up: complete
Overall study quality: low risk of bias
ITT: yes
Sample size calculation reported


ParticipantsPremature infants with respiratory distress syndrome (RDS) in the neonatal ICU, 942 screened with birth weights 750-1900 grams.

AT III: 61 patients. 33 male, gestational age (weeks) 28.3 ±2.03, birth weight: 1.198±300.9 g, venous cord blood pH: 7.31±0.10, APGAR score at 1 min: 5.4±1.95, APGAR score at 5 min: 7.6±1.24, AT III, U/mL: 0.33± 0.08, maternal age: 28.1± 6.68, maternal steroids: 40 (65.6%).

Control: 61 patients. 31 male, gestational age (weeks) 28.8 ±2.25, birth weight: 1.201±314.4 g, venous cord blood pH: 7.28±0.15, APGAR score at 1 min: 5.3±2.22, APGAR score at 5 min: 7.7±1.16, AT III, U/mL: 0.32± 0.08, maternal age: 28.9± 5.16, maternal steroids: 35 (57.4%).

Inclusion criteria: 224 met inclusion criteria; (1) age between 2-12 h, 2) endotracheal intubation and positive pressure ventilation for RDS, 3) indwelling arterial catheter, 4) ratio of arterial to alveolar oxygen pressure (a/A) PO2 <0.3 after the first dose of exogenous surfactant).

Exclusion criteria: 1) congenital infection, 2) congenital malformation(s), 3) hydrops, 4) pulmonary hypoplasia, 5) clinically apparent bleeding disorder, 6) thrombocytopenia (platelets =/< 50 x 109/L), 7) moribund.


InterventionsAT III: loading dose 100 U/kg followed by 50 U/kg every 6 h for 48 h.

Control: human albumin 1%, equivalent dose.

All participants received the same standard intensive care treatment and the same exogenous surfactant. Arterial catheters were perfused with 1 ml/h of 5% dextrose containing 1 IU/ml of unfractionated heparin. No additional administration of anticoagulants.


OutcomesPrimary: mortality.

Secondary: days in ICU, plasma AT activity, TAT-complex (thrombin-AT), prothrombin fragment (F1+2), ratio of arterial to alveolar oxygen pressure (a/A) PO2 and ventilator efficiency index (VEI), thrombin formation, Improved gas exchange, duration of mechanical ventilation, duration of supplemental oxygen therapy, safety of AT therapy evaluated primarily by comparing the incidence of severe (Grade 3) intraventricular haemorrhage and of periventricular echodensities in the 2 groups, bleeding events.


NotesCountry: Canada
Letter sent to authors in November 2005, no answer received.
AT levels in healthy preterm infants are about 40% of normal adult values, and even lower levels have been reported in sick premature infants with RDS. The therapy was initiated on average 7.2 h after delivery in each group.
Having observed for some time an apparent imbalance in deaths between the treatment groups, the external safety and efficacy monitoring committee decided to break the code and make its recommendation because 7 deaths had occurred in patients receiving AT and two with placebo.

Authors' conclusion: AT III therapy cannot be recommended in premature infants with RDS. AT III did not improve gas exchange in premature infants with RDS. It may even be harmful. The durations of mechanical ventilation and oxygen therapy were significantly prolonged, whereas deaths, periventricular and intraventricular haemorrhages and clinically apparent bleeding episodes all tended to be more common in infants randomized to AT concentrate.


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Low riskAdequate

Allocation concealment (selection bias)Low riskA - Adequate

Schorr 2000

MethodsTwo-group parallel randomized clinical trial
Generation of allocation sequence: unclear
Allocation concealment: unclear
Blinding: no blinding
Follow up: adequate
Overall study quality: high risk of bias
ITT: yes
Sample size calculation was not reported


ParticipantsPatients with secondary peritonitis, surgical population in ICU.

AT III: 24 patients, 12 men, median age 60 (33-86) y, APACHE II 14.4 (6-30), MOF 4.6 (0-10), SOFA 7.7 (0-18), Mannheimer peritonitis index (MPI) 26.7 (12-38), malignant diseases (%) 33. Origin of peritonitis: a) large bowel 14, b) small bowel 7, c) gastric 3, d) biliary system 1, uterus 0.

Control: 26 patients, 12 men, median age: 62 (28-85) y, APACHE II 14.6 (4-28), MOF 4.9 (0-11), SOFA 8.7 (0-16), MPI 27.8 (16-38), malignant diseases (%) 35. Origin of peritonitis: a) large bowel 13, b) small bowel 6, c) gastric 4, d) biliary system 1, uterus 1.

Inclusion criteria: 1) > 18 years, 2) diffuse secondary peritonitis affecting two or more quadrants after perforation of an intraperitoneal organ or failed anastomosis, 3) focus of infection had to be surgically eliminated.

Exclusion criteria: 1) acute pancreatitis, 2) peritonitis due to peritoneal dialysis, 3) primary peritonitis, 4) ascites due to cirrhosis of the liver or due to malignant underlying disease, 5) pregnancy, 6) incapacitated persons, 7) prisoners.


InterventionsAT III: continuous IV AT III and 2 intraperitoneal installations of fresh frozen serum (FFS). The aim was to achieve AT III level of at least 140% of the normal plasma value for 4 days. A calculated bolus was given in 1 hour followed by a continuous infusion of AT III (200 IU-800 IU per hour) depending on the 6-hourly measurements of plasma AT III. 300 ml of blood group compatible FFS was applied intraperitoneally after the operation and 6 hours post-operatively through especially installed drains. The first FFS was supplemented with 1500 IU AT III to equalize the lack of AT III in FFS. Mean AT III administered was 26196 (±299 SEM) IU.

Control: no placebo.

All patients: routine intensive care treatment. If no risk of postoperative bleeding, heparin iv 200-400 IU/h 6 hours after operation until discharge from ICU.


OutcomesPrimary: 90 day mortality.

Secondary: number of days at ICU, mechanical ventilation, organ function scores, side effects, organ failure and function, bleeding and transfusions, AT III activity, prothrombin concentrations in exudate, TAT-complex in exudate, opsonic capacity and opsonin concentration in exudate.


NotesCountry: Germany
Letter sent to authors in November 2005, no reply received.

Authors' conclusion: the chosen therapeutic approach was feasible and showed no side effects. Mortality nor multiple organ failure were significantly improved by the applied short-term adjuvant therapy.


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Unclear riskUnclear

Allocation concealment (selection bias)Unclear riskB - Unclear

Smith-Erichsen 1996

MethodsTwo-group parallel randomized clinical trial
Generation of allocation sequence: unclear
Allocation concealment: unclear
Blinding: unclear
Follow up: adequate, until discharge from hospital or death
ITT: no
Overall study quality: high risk of bias
No sample size calculation was reported


ParticipantsCritically ill and trauma patients in ICU.
83 patients allocated to intervention and control groups. Within each of these groups, the patients again were allocated to 2 groups (group 1 and group 2) according to the type of surgery or trauma.

Group 1: multiple trauma defined as injury of at least 2 body regions, but without abdominal injury.

Group 2: abdominal operations and multiple trauma with abdominal injury; 2.1) total gastrectomy, oesophageal resection, rectal resection/amputation in patients over 70 years old, 2.2) pancreatic resection, 2.3) liver resection, 2.4) secondary gastro-intestinal interventions due to post-operative complications, 2.5) multiple trauma with abdominal injury. The patients were randomized in blocks with 4 patients in each block.

AT III group: 43 patients, 30 men, age: 60.1±16.3 y. 5 patients allocated to group 1 (trauma). Number of patients in group 2 as defined above; Group 2.1=17, Group 2.2=9, Group 2.3=5, Group 2.4=6, Group 2.5=1.

Control group: 40 patients, 25 men, age: 59.0±16.2 y. 5 patients allocated to group 1 (trauma). Number of patients in group 2 as defined above; Group 2.1=17, Group 2.2=8, Group 2.3=4, Group 2.4=5, Group 2.5=1.

Inclusion criteria: AT III plasma level less than 80% measured the day after trauma or major abdominal surgery.

Exclusion criteria: trauma patients with head injury and patients below the age of 18 years.


InterventionsAT III: the aim was to maintain plasma AT III activity in the range of 100%±10%. No data were provided on average or total dose. Based on daily samples, the substitution dose was calculated. A maintenance dose of 500 IU was given at midnight and 10 hours the following day. Treatment was continued until the plasma AT III activity was maintained within the normal range for 3 consecutive days without AT III substitution but was discontinued after 14 days, irrespective of the plasma activity.

Control: no placebo.

All patients were treated according to the standard management routines of each hospital.

500 ml of dextran-70 was given every 2nd day during the 1st week as antithrombotic prophylaxis. No patient received heparin or low-molecular weight heparin during the observations period


OutcomesPrimary: plasma protease changes, mortality, days in ICU, days in hospital
Secondary: AT III level, prothrombin, prekallikrein, functional kallikrein inhibition, plasma kallikrein-like activity, plasminogen, functional antiplasmin activity and plasmin-like activity, proenzyme functional inhibition index and adverse effects.


NotesCountry: Norway
No adverse effects were reported.
Letter sent to authors in December 2006, no answer received.

Authors' conclusion: In this study both major surgery and trauma induced changes in components of the plasma protease systems. Substitution therapy with AT III concentrate to normalise plasma levels of this antiprotease did not modify these changes.


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Unclear riskUnclear

Allocation concealment (selection bias)Unclear riskB - Unclear

Warren 2001

MethodsTwo-group parallel randomized multicenter clinical trial
Generation of allocation sequence: adequate
Allocation concealment: adequate
Blinding: adequate
Follow up: inadequate
Overall study quality: low risk of bias
ITT: yes
Sample size calculation reported


ParticipantsCritically ill population of ICU patients with severe sepsis and septic shock.

AT III: 1157 patients, men 62%; age mean (SD) = 57 (17) years; SAPS II (SD) = 49 (17); body weight mean (SD) = 77 (19) kg. Underlying problem or site of infection: 1) respiratory system 35%, 2) intra-abdominal infection 27%, 3) genitourinary system 6%, 4) injury 7%, 5) other 24%. Surgical status: A) no 54%, yes 46%. Circulatory shock: 49%; blood culture results: a) gram-negative 15%, b) gram-positive 15%, c) other/mixed 4%, d) not done/not verified 66%.

Control: 1157 patients, men 61%; age mean (SD) = 58 (17) years; SAPS II (SD) = 49 (16); body weight mean (SD) = 77 (20) kg. Underlying problem or site of infection: 1) respiratory system 34%, 2) intra-abdominal infection 28%, 3) genitourinary system 8%, 4) injury 6%, 5) other 24%. Surgical status: A) no 53%, yes 47%. Circulatory shock: 47%; blood culture results: a) gram-negative 16%, b) gram-positive 17%, c) other/mixed 2%, d) not done/not verified 64%.

Inclusion criteria: 18 years or above, and met following criteria within 6 hours: 1) clinical evidence of sepsis with a suspected source of infection, 2) body temperature (rectal or core) >38.5 ºC or <35.5 ºC, 3) leukocyte count >10,000/µL or <3500/µL. Additionally, 3 of the following 6 signs had to be met within the same 6-hour period: A) heart rate >100/min, B) tachypnoea >24/min or mechanical ventilation because of septic indication, C) hypotension, systolic BP <90 mm Hg despite sufficient fluid replacement or the need of vasoactive agents to maintain systolic BP of 90 mm Hg or greater, D) platelet count <100,000/µL, E) elevated lactate above upper limits of normal range or metabolic acidosis (pH <7.3 or BE</= -10 mmol/L) not secondary to respiratory alkalosis, F) oliguria (<20ml/h despite sufficient fluid replacement).

Exclusion criteria: a) advanced directive to withhold life-sustaining treatment (except cardiopulmonary resuscitation), b) condition other than sepsis anticipated to be fatal within 28 days, c) pregnancy or breast feeding, d) history of hypersensitivity to study medication, e) treatment with other investigational drugs within the last 30 days, f) treatment with an AT III concentrate within the last 48 hours, g) treatment with heparin (except sc low dose or iv line flushing) or coumarin derivatives, h) NSAID treatment within previous 2 days, i) known bleeding disorder or ongoing massive surgical bleeding, j) platelet count <30x1000/µL, k) immunocompromised status, l) acute myocardial infarction (within previous 7 days), m) third-degree burns =/ >20% of total body area, n) incurable malignancy with metastases and life expectancy of <3 months, o) haematologic neoplasia during cytostatic treatment, p) bone marrow aplasia, q) pre-existing dialysis-dependent renal failure, r) end-stage liver disease, s) transplantation (post-operative state), t) history of stroke within the last year, u) severe cranial or spinal trauma within the last year, v) planned cranial or spinal surgery (except nontraumatic lumbar puncture) within the next 48 hours.


InterventionsAT III: loading dose of 6000 IU given over 30 minutes, followed by a continuous iv infusion of 6000 IU per day for 4 days, total of 30000 IU.

Control: equivalent volume of placebo solution (1% of human albumin).

Heparin was permitted to be used in prophylactic doses as an adjunct to standard therapy. Administration of heparin was left to the attending physicians. Not all patients received heparin (698 patients received no heparin while receiving AT). Heparin dosage: unfractionated or low molecular weight heparin for venous thrombosis prophylaxis (</= 1000 IU subcutaneous per day) and heparin flushes for vascular catheter patency (iv of </= 2I U/kg body weight/h).


OutcomesPrimary: all-cause mortality at 28 days (subgroup 28 and 90-day survival for patients not receiving heparin).

Secondary: mortality at 56 and 90 days, survival time, length of ICU stay, occurrence of new organ dysfunction within 7 days (according to logistic organ dysfunction score), severity of sepsis (SAPSII), circulatory shock index (ie ratio of heart rate (beats/minute) and systolic BP (mm Hg) exceeded the value of 1.5); surgical interventions and bleeding events for 28 days, major bleeding if intracranial or required transfusion of >3 units of blood, other serious adverse events, AT III plasma concentration, activated partial thromboplastin time and prothrombin time.


NotesCountry: KyberSept Trial Study Group, 211 contributing centres in 19 countries including USA, Germany

Letter sent to authors November 2005, reply by authors in December 2005 and January 2006.

Authors' conclusion: high-dose AT III therapy had no effect on 28-day all-cause mortality in adult patients with severe sepsis and septic shock when administered within 6 hours after the onset. There was no statistically significant difference either in the 56 or 90-day mortality. High-dose AT III was associated with an increased risk of haemorrhage when administered with heparin. Treatment benefit became evident in the subgroup of patients not receiving concomitant heparin and the longer the observation period, the greater was the difference (on day 90 it reached statistical significance).


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Low riskAdequate

Allocation concealment (selection bias)Low riskA - Adequate

Waydhas 1998

MethodsTwo-group parallel randomized clinical trial
Generation of allocation sequence: inadequate
Allocation concealment: adequate
Blinding: adequate
Follow up: adequate
Overall study quality: high risk of bias
ITT: yes
Sample size calculation not reported


ParticipantsTrauma patients in ICU.

AT III: 20 patients, 13 men, age 32 (26-42) y, systolic blood pressure (SBP) on the scene 120 mm Hg (90-130), 2 patients with SBP <80 mm Hg, Glasgow coma score 13 (6-15), use of catecholamines on the scene in 2 cases, intubation on the scene 15 (75%), fluids on the scene 1.5 L (1.0-3.0), time from accident until admission 53 min (33-75), lactate 3.2 mmol/L (2.7-8.0), haemoglobin 8.9 g/100 mL (7.0-12-0), PH 7.30 (7.27-7.36), base deficit 6.2 (8.3-3.7).

Control: 20 patients, 13 men, age 35 (27-47) y, SBP on the scene 100 mm Hg (93-120), 2 patients on the scene with SBP <80 mm Hg, Glasgow coma score 13 (4-15), use of catecholamines on the scene in 2 cases, intubation on the scene 10 (50%), fluids on the scene 2.0 L (1.0-3.0), time from accident until admission 58 min (30-105), lactate 3.7 mmol/L (2.1-6.0), haemoglobin 9.5 g/100 mL (8.0-11-0), PH 7.29 (7.10-7.32), base deficit 7.6 (13.1-4.2).

Inclusion criteria: 1) injury severity score (ISS) of 29 or greater (because an ISS of 29 or greater was associated with an organ failure rate of more than 60%), 2) admission within 6 h after injury, 3) 18-70 years, 4) blunt trauma, 5) if the head region contributed to calculation of the ISS, the sum of squared abbreviated ISS of the 2 other regions that were included in the calculation had to be equal or greater than 20 points in order to rule out predominant head injury and ensure severe multiple injuries.

Exclusion criteria: a) resuscitation with epinephrine, (b) if closed chest compression performed, c) severe brain injury with uncertain outcome (both pupils unresponsive to light for more than 30 minutes, midline shift >1 cm, or extensive intracerebral haemorrhage on the initial computed tomographic scan).
The only post hoc exclusion criterion was survival of less than 24 hours after trauma.


InterventionsAT III: a total of 20.000 IU (16.125-22.875), vials containing 500 IU diluted in 8 mL of water. Infusion-rate of 96mL/h via a pump. AT III levels assessed every 6 h during the first 48 h. Additional AT III or placebo was substituted to keep the AT III concentration at 140% of normal. In addition, on the next 2 days the test substance was administered once daily in the morning.

Control: placebo, 20% human albumin in corresponding doses and volume.

For prophylaxis of DVT and pulmonary embolism: standard sodium heparin iv at a rate of 400 IU/h, started within the first 24 h after trauma unless contraindications (active bleeding, intracranial haemorrhage, or others). Patients with traumatic intracranial bleeding were started on heparin on day 5, when the acute phase of brain swelling had subsided and there were no signs of ongoing haemorrhage.


OutcomesPrimary: incidence and severity of multiple organ dysfunction, mortality, incidence of respiratory failure, severity of organ failure, duration of mechanical ventilation and length of stay in the ICU and hospital.

Secondary: plasma concentration of indicators of DIC and systemic inflammatory response (prothrombin, prothrombin fragment F1+F2, TAT-complex lower, partial thromboplastin time, prothrombin time, PC, plasminogen activator inhibitor I, soluble TNF receptor II, IL1 receptor antagonist, IL6, IL8, neutrophil elastase), transfusions, respiratory dysfunction, ARDS, renal and liver dysfunction, incidence of DIC.


NotesCountry: Germany
Letters sent to authors in November and December 2005, authors replied both times.

Authors' conclusion: the early and high-dose administration of AT III to patients with severe blunt trauma appears not to extenuate the posttraumatic inflammatory response or to significantly improve outcome.

8 patients were excluded due to survival less than 24 hours; 3 in control group and 5 in AT group due to severe brain injury (n = 5) and uncontrollable haemorrhage (n = 3). During the first 48 hours after trauma, 12 and 9 patients in the control and AT III groups, respectively, received heparin prophylaxis. At day 6, all but one patient in each group was receiving heparin. The sample size was too small for a reliable subgroup analysis bases on heparin administration.


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)High riskInadequate

Allocation concealment (selection bias)High riskC - Inadequate

 
Characteristics of excluded studies [ordered by study ID]

StudyReason for exclusion

Balk 1995Septic patients randomized to receive either AT III or placebo in a double-blind treatment protocol. 34 patients followed for 28 days with organ dysfunction/failure evaluation. The overall mortality was 22.2% in AT III group and 12.5% in placebo. Reason for exclusion: published only as an abstract.

Banfi 1996An observational study on the use of AT III in 16 ICUs in Italy with no reported reduction in mortality or in the length of stay in hospital. Reason for exclusion: no randomization.

Blauhut 19853-group parallel randomized clinical trial comparing effects of AT III and heparin, AT III without heparin and heparin alone. Heparin concentration was different in the AT III + heparin group than the heparin group. Reason for exclusion: multiple treatments.

Brangenberg 1997103 preterm infants at a gestational age of 25-32 weeks (mean 28.9 weeks; birth weight 600-2170 g, mean 1285 g) received AT III at a single dosage of 50-200 IU/kg on the day of birth and subsequently only in case of a new fall in AT III activity below 50%. Coagulation parameters, incidence of intraventricular haemorrhage (IVH) and other clinical outcomes were reported. Reason for exclusion: observational study, no control group, no randomization.

Danielsson 1997The study was divided into 2 parts with a total of 14 consecutive patients (per cent total body surface area after burn injury > or = 20%). The first 6 patients were given AT III when AT III fell below 50%. The second part examined the restitution of the coagulation parameters when the patients (n = 8) obtained AT III substitution only at extremely low values of AT III. Reason for exclusion: observational study with no randomization or control group.

du Cheyron 2006Observational study evaluating the effect of AT III supplementation on filter lifespan compared with heparin in critically ill patients with septic shock requiring continuous renal replacement therapy (CRRT). AT III supplementation was independently associated with a decrease in clotting rate. Mortality did not differ. Reason for exclusion: retrospective case-control analysis based on a 4-year observational study. No randomization.

Fertmann 2006This study examined prophylactic high-dose AT III application in patients with first pancreas-kidney (SPK) transplantation. 53 consecutive patients with SPK were studied. In AT III group (n=24) 3000 IU of AT III was given intravenously before pancreatic reperfusion, while the control group (n=29) received standard therapy including post-operative AT III supplementation. Reason for exclusion: no randomization, retrospective analysis.

Fourrier 1990Effect of early infusion of AT III among 5 adult patients with meningococcaemia (type B) and purpura fulminans. Designed to measure the levels of AT III, protein C, protein S and survival. Reason for exclusion: no randomization, case report.

Hanada 1985Effect of AT III bolus injections (250 U/8hours) and/or continuous heparin infusion (400 U/kg/day) in 6 children with DIC (1 month to 5 years old). Reason for exclusion: no randomization, case report.

Haussmann 2006This trial comprised of two phases: 1) 71 children did not receive any specific veno-occlusive disease (VOD) prophylaxis or therapy (controls); 2) 91 children were given pre-emptive AT III replacement in case of decreased AT III activity (< or =70%). If VOD was diagnosed clinically, high-dose defibrotide (60 mg/day) and AT III replacement therapy were combined.
Reason for exclusion: no randomization, case series.

Hellgren 1984A study of AT III supplementation in 9 severely ill patients with DIC. Reason for exclusion: no randomization, no control group.

Hoffmann 2004This paper reports the effects of AT III supplementation on septic coagulatory response in patients with severe sepsis when given over 14 days. Reason for exclusion: based on the same trial as Inthorn 1997, post hoc analysis.

Hoffmann 2006This paper analyses heparin-AT III interactions in terms of long-term mortality, adverse events, and thromboembolic events. Reason for exclusion: based on the same trial as Warren 2001, post hoc analysis.

Ilias 2000An open randomized clinical trial, evaluating the safety, pharmacokinetics and practicability of two different regimens of AT III treatment (intermittent bolus infusions vs continuous infusion), both aiming at a total dose of 30,000 IU AT III with concomitant heparin during a 4-day period. Reason for exclusion: two different regimens of AT III treatment.

Inthorn 1998This paper reports the effects of long-term AT III supplementation on the inflammatory response in a septic population. Reason for exclusion: based on the same trial as Inthorn 1997.

Jochum 1995This review includes one unpublished, not double-blinded trial of AT III supplementation in a septic population aiming at AT III activity > 120%. AT III group (n = 20) received 2000-8000 U/d AT III over 21 days and heparin (4 U/kg body weight/hour). Control group (n=20) received no placebo. Outcome measures were various organ dysfunction analyses, inflammatory and coagulatory parameters and mortality (control: 80%, AT III: 65%). The results were not statistically significant. Reason for exclusion: review and data from an unpublished trial.

Kienast 2006This paper examines the treatment effects of high-dose AT III in patients with severe sepsis with or without DIC. 563 patients from Warren 2001 were identified (placebo 277; AT III 286) who did not receive concomitant heparin and had sufficient data for DIC determination. Reason for exclusion: based on the same trial as Warren 2001, post hoc analysis.

Korninger 1987A randomized trial of AT III in patients undergoing peritoneo-venous shunt operation because of intractable ascites. 10 patients with alcoholic liver cirrhosis were randomized according to the stage of liver disease and pre-operative AT III levels. AT III was infused in 5 patients, twice daily for 4 days, at a dose of 20 U/kg body weight starting 12 hours prior to operation. 5 patients were allocated to placebo. Coagulatory parameters were examined. Reason for exclusion: only published as abstract, no data on mortality.

Kowal-Vern 2000In this trial 17 patients with thermal injuries of a second or third-degree burn covering more than 20% of body surface either received AT III supplementation or standard therapy. Outcome measures were wound healing, estimated blood loss, scarring, level of contractures, and length of stay. Reason for exclusion: allocation to the groups at the patients choice, no randomization.

Leitner 2006This randomized, double-blinded, placebo-controlled study examined 30 healthy male volunteers. The active treatment groups received infusions of AT III to achieve levels of 200% and 500% before infusion of 2 ng/kg endotoxin (LPS). Reason for exclusion: experimental trial among healthy volunteers.

Maki 1987This randomized controlled trial compared two active interventions, being AT III with FOY (injectable protease inhibitor: gabexate mesilate) in patients with obstetric complications. Reason for exclusion: two different active interventions.

Messori 2002Observational retrospective study of 216 critically ill patients treated with AT III at 20 Italian hospitals. Reason for exclusion: retrospective, no randomization, no control group.

Mitchell 2003bIn this study children with ALL were screened for thrombotic events. Reason for exclusion: part of the PARKAA study (Mitchell 2003); only data on children without AT III treatment were provided.

Moriss 1997An observational study of 10 patients with severe post-bone marrow transplantation organ dysfunction (single or multiorgan) and AT III levels < 88% were treated with AT III concentrate. All patients were loaded with 50 units/kg AT III every 8 h for 3 doses followed by 50 units/kg/day each day for 3-12 days. Reason for exclusion: no randomization, no control group.

Muntean 1989This open, controlled, randomized trial examined a single administration of AT III concentrate in 103 premature infants (gestational age 26-34 weeks, mean 31). No placebo. No benefit of AT III administration was found. Trend towards more frequent deaths in the AT III group (9/45) compared with control group (8/53). Reason for exclusion: only published as an abstract.

Nishiyama 2006This randomized, unblinded trial investigated the effects of AT III on coagulation, fibrinolysis, production of cytokines and adhesion molecules in abdominal aortic aneurysm repair surgery. 16 patients for Y-shaped graft replacement of abdominal aortic aneurysm were divided into an AT III group and a control group. In the AT III group, 3000 U AT III was infused over 30 min before heparin administration and 24 h later. Reason for exclusion: no data on mortality, nor any data relevant for our review.

Palareti 1995In this trial, 119 patients (age 18-75 y, ATIII < 70% with sepsis and/or post-surgical complications requiring haemodynamic or respiratory support) were randomized to receive AT III (4000 U as bolus + 2000 U/12 hours by continuous infusion for 5 days) or placebo. Laboratory effects of AT III replacement were examined. Reason for exclusion: only published as an abstract.

Paternoster 2000This prospective study evaluated the modifications of clotting and clinical parameters before and immediately after delivery amongst a pre-eclamptic population either treated with AT III (n=18) or acting as control (n=21). Reason for exclusion: only published as a brief communication, no data on randomization or blinding.

Paternoster 2004In this study, 23 pre-eclamptic women were randomly subdivided into 2 groups: first group (n=10) were treated with 3000 units AT III once daily for 5 days or until delivery, while the second group (n=13) were treated with doses of AT III sufficient to maintain at least 80% of the activity. The end points were the prolongation of pregnancy, assessment of the maternal bleeding and haemostasis, and inflammatory markers. Reason for exclusion: comparison of 2 different regimens of AT III.

Scherer 1994In this trial, 51 adult patients scheduled for orthotopic liver transplantation were divided into 2 groups. In 37 patients with AT III activity < 70%, AT III was given pre-operatively (mean dose 2616 +/- 207 IU) in order to increase AT III activity to 80%. In the other group,14 patients with prothrombin time (PT) < 40%, and with a pre-corrected AT III during the clinical course (AT III activity > 70%), prothrombin complex was substituted (mean dose 2304 +/- 289 IU) aiming at a PT > 60%. Reason for exclusion: no randomization, 2 different interventions.

Scherer 1997A randomized prospective study examining the effects of AT III substitution in order to achieve supranormal values in patients with liver cirrhosis scheduled for liver transplantation.
19 patients were given AT III aiming at either 100% (n = 10) or 175% (n = 9) AT III activity. Control group (n = 5) received saline 0.9%. The end points were various coagulatory variables measured prior to AT III infusion and 60 min thereafter. Reason for exclusion: no data on mortality, 3 groups.

Schuster 1997A study of 45 patients with sepsis without DIC, allocated to AT III supplementation (loading dose of 3000 IU followed by 500 IU every 4 hrs for 7 days, total dose 17,000 IU) or placebo (albumin). Both groups received iv heparin (6 IU/kg/h). Reported mortality was: AT III 9/22, control 10/23. Reason for exclusion: only submitted as an abstract.

Seitz 1989An open pilot study of 42 patients with septic shock. AT III group (n= 29) treated with AT III, fresh-frozen plasma, and continuous heparin infusion at an average dose of 300 U/h. Control group (n = 13) received similar heparin regimen. Reason for exclusion: no randomization.

Shimada 1994A randomized prospective trial of AT III efficacy in hepatic resection among 24 patients with hepatocellular carcinoma. 13 patients were given 1500 IU AT III immediately before operation, during hepatectomy, and immediately after operation. Control group (n = 11) received no placebo. Coagulant and fibrinolytic profiles were determined. Reason for exclusion: no data on mortality, different population than defined in our review.

Terao 1989A randomized trial of 40 patients with pre-eclampsia comparing AT III intervention with no treatment on coagulatory findings, gestosis Index, abortion times, and incidence of newborn asphyxia. Reason for exclusion: no data on mortality.

van Beek 1994A study of adults with sepsis and AT III levels < 0.45 IU/ml. AT III group (n= 21) received AT III bolus of 20 IU/kg followed by continuous infusion of 20 IU/g per 24 h. Control (n= 21) received no placebo. Mortality: AT III group 12/21 (95% CI 34-78%), control 16/21 (95% CI 53-92%). Reason for exclusion: retrospective, no randomization.

van Leeuwen 2003This paper examined the changes in lipoprotein levels in patients with severe sepsis. Reason for exclusion: different objective, based on the same trial as Warren 2001, post hoc analysis.

Vinazzer 1995This review mentioned one unpublished trial of patients with DIC due to sepsis or septic shock in which 170 patients were evenly allocated to two groups (heparin versus AT III). This uncontrolled open clinical trial showed positive outcome of AT III substitution. Reason for exclusion: review and data from an unpublished trial.

Von Kries 1985A trial of AT III infusion after prior activation with heparin among 10 newborns with DIC. Primary outcome: AT III activity. Reason for exclusion: no randomization, case report.

Wiedermann 2006Re-analysis of prospectively defined populations from the KyberSept trial. Reason for exclusion: based on the same trial as Warren 2001, post hoc analysis.

 
Characteristics of ongoing studies [ordered by study ID]
D'angelo 2005

Trial name or titleEPAS (Early Preeclampsia Antithrombin Study)

Methods

ParticipantsWomen with pre-eclampsia occurring before the 30th week of gestation.

Interventions7-day course of antithrombin 3000 IU/day. No heparin prophylaxis will be permitted during the time of AT III administration.

OutcomesPrimary outcome: reduction of the combined end point of fetoneonatal mortality and severe neonatal morbidity.
Secondary outcomes: reduction in maternal complications and stay in neonatal intensive care unit and the evaluation of the changes in coagulation and inflammation parameters.

Starting dateDecember 2004

Contact informationD'angelo A (1),Valsecchi L (2)
1): Coagulation Service and Thrombosis Research Unit
2): Department of Obstetrics and Gynecology, Irccs
H S. Raffaele Milano, Italy

NotesD'angelo A,Valsecchi L; A double-blind study of high-dose Antithrombin supplementation in early preeclampsia. Thrombosis Research. 2005; 115 Suppl 1:117

 
Comparison 1. AT III versus control

Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size

 1 Mortality (subgroup analysis on bias risk)203458Risk Ratio (M-H, Fixed, 95% CI)0.96 [0.89, 1.03]

    1.1 Trials with low bias risk
82875Risk Ratio (M-H, Fixed, 95% CI)0.95 [0.88, 1.03]

    1.2 Trials with high bias risk
12583Risk Ratio (M-H, Fixed, 95% CI)1.00 [0.78, 1.27]

 2 Overall mortality (subgroup analysis on random sequence generation)203458Risk Ratio (M-H, Fixed, 95% CI)0.96 [0.89, 1.03]

    2.1 adequate random sequence generation
92957Risk Ratio (M-H, Fixed, 95% CI)0.95 [0.88, 1.03]

    2.2 Inadequate or unclear random sequence generation
11501Risk Ratio (M-H, Fixed, 95% CI)0.99 [0.78, 1.25]

 3 Overall mortality (subgroup analysis on allocation concealment)203458Risk Ratio (M-H, Fixed, 95% CI)0.96 [0.89, 1.03]

    3.1 adequate allocation concealment
123060Risk Ratio (M-H, Fixed, 95% CI)0.96 [0.88, 1.04]

    3.2 Inadequate or unclear allocation concealment
8398Risk Ratio (M-H, Fixed, 95% CI)0.94 [0.72, 1.21]

 4 Overall mortality (subgroup analysis on blinding)203458Risk Ratio (M-H, Fixed, 95% CI)0.96 [0.89, 1.03]

    4.1 blinded
102957Risk Ratio (M-H, Fixed, 95% CI)0.95 [0.88, 1.03]

    4.2 Not blinded unclear blinding
10501Risk Ratio (M-H, Fixed, 95% CI)0.99 [0.76, 1.28]

 5 Overall mortality (subgroup analysis on completeness of follow up)203458Risk Ratio (M-H, Fixed, 95% CI)0.96 [0.89, 1.03]

    5.1 Absence of complete follow up
82855Risk Ratio (M-H, Fixed, 95% CI)0.95 [0.88, 1.04]

    5.2 Complete follow up
12603Risk Ratio (M-H, Fixed, 95% CI)0.97 [0.77, 1.23]

 6 Overall mortality (subgroup analysis on ITT)203458Risk Ratio (M-H, Fixed, 95% CI)0.96 [0.89, 1.03]

    6.1 ITT
13814Risk Ratio (M-H, Fixed, 95% CI)0.96 [0.77, 1.19]

    6.2 No ITT
72644Risk Ratio (M-H, Fixed, 95% CI)0.96 [0.88, 1.04]

 7 Overall mortality (subgroup analysis on median follow up)203458Risk Ratio (M-H, Fixed, 95% CI)0.96 [0.89, 1.03]

    7.1 Mortality in trials with follow up less than median of all trials
11679Risk Ratio (M-H, Fixed, 95% CI)0.96 [0.76, 1.20]

    7.2 Mortality in trials with follow up longer than median of all trials
92779Risk Ratio (M-H, Fixed, 95% CI)0.96 [0.88, 1.04]

 8 Overall mortality (subgroup analysis on duration of intervention)203458Risk Ratio (M-H, Fixed, 95% CI)0.96 [0.89, 1.03]

    8.1 Median duration of AT III intervention equal to or less than one week
173250Risk Ratio (M-H, Fixed, 95% CI)0.96 [0.89, 1.04]

    8.2 Median duration of AT III intervention longer than one week
3208Risk Ratio (M-H, Fixed, 95% CI)0.89 [0.59, 1.34]

 9 Overall mortality (trauma)268Risk Ratio (M-H, Fixed, 95% CI)2.15 [0.81, 5.72]

    9.1 Trials with high bias risk
268Risk Ratio (M-H, Fixed, 95% CI)2.15 [0.81, 5.72]

 10 Overall mortality (obstetrics)2333Risk Ratio (M-H, Fixed, 95% CI)1.03 [0.33, 3.21]

    10.1 Overall maternal mortality, trials with low bias risk
2175Risk Ratio (M-H, Fixed, 95% CI)0.0 [0.0, 0.0]

    10.2 Overall fetal and neonatal mortality, trials with low bias risk
2158Risk Ratio (M-H, Fixed, 95% CI)1.03 [0.33, 3.21]

 11 Overall mortality (paediatrics)3267Risk Ratio (M-H, Random, 95% CI)1.49 [0.38, 5.87]

    11.1 Trials with low bias risk
2182Risk Ratio (M-H, Random, 95% CI)1.49 [0.38, 5.87]

    11.2 Trials with high bias risk
185Risk Ratio (M-H, Random, 95% CI)0.0 [0.0, 0.0]

 12 Overall mortality (heparin; Warren 2001 as a trial with adjuvant heparin therapy)203458Risk Ratio (M-H, Fixed, 95% CI)0.96 [0.89, 1.03]

    12.1 Trials with complete or partially adjuvant heparin therapy
143033Risk Ratio (M-H, Fixed, 95% CI)0.96 [0.89, 1.04]

    12.2 Trials without adjuvant heparin
6425Risk Ratio (M-H, Fixed, 95% CI)0.92 [0.67, 1.27]

 13 Overall mortality (heparin; Warren 2001 as a trial without adjuvant heparin therapy)203458Risk Ratio (M-H, Fixed, 95% CI)0.96 [0.89, 1.03]

    13.1 Trials with complete or partially adjuvant heparin therapy
13719Risk Ratio (M-H, Fixed, 95% CI)0.98 [0.79, 1.22]

    13.2 Trials without adjuvant heparin
72739Risk Ratio (M-H, Fixed, 95% CI)0.95 [0.88, 1.04]

 14 Overall mortality (heparin; Warren 2001 data split based on Heparin administration)203458Risk Ratio (M-H, Random, 95% CI)0.95 [0.88, 1.03]

    14.1 Trials with complete or partially adjuvant heparin therapy
142335Risk Ratio (M-H, Random, 95% CI)0.99 [0.90, 1.09]

    14.2 Trials without adjuvant heparin
71123Risk Ratio (M-H, Random, 95% CI)0.89 [0.77, 1.02]

 15 Complications during the in-patient stay specific to the trial intervention (intracranial bleeding)32454Risk Ratio (M-H, Random, 95% CI)1.26 [0.83, 1.92]

    15.1 Trials with low bias risk
22429Risk Ratio (M-H, Random, 95% CI)1.62 [0.96, 2.73]

    15.2 Trials with high bias risk
125Risk Ratio (M-H, Random, 95% CI)0.92 [0.51, 1.66]

 16 Complications during the in-patient stay not specific to the trial intervention (renal failure)265Risk Ratio (M-H, Random, 95% CI)0.71 [0.08, 6.11]

    16.1 Trials with high bias risk
265Risk Ratio (M-H, Random, 95% CI)0.71 [0.08, 6.11]

 17 Complications during the in-patient stay not specific to the trial intervention (pneumothorax)160Risk Ratio (M-H, Fixed, 95% CI)0.67 [0.12, 3.71]

    17.1 Trials with low bias risk
160Risk Ratio (M-H, Fixed, 95% CI)0.67 [0.12, 3.71]

 18 Complication specific to the trial intervention other than bleeding3187Risk Ratio (M-H, Fixed, 95% CI)0.72 [0.42, 1.25]

    18.1 Trials with low bias risk
160Risk Ratio (M-H, Fixed, 95% CI)0.75 [0.18, 3.07]

    18.2 Trials with high bias risk
2127Risk Ratio (M-H, Fixed, 95% CI)0.72 [0.40, 1.30]

 19 Bleeding events92929Risk Ratio (M-H, Random, 95% CI)1.52 [1.30, 1.78]

    19.1 Trials with low bias risk
62791Risk Ratio (M-H, Random, 95% CI)1.37 [1.04, 1.82]

    19.2 Trials with high bias risk
3138Risk Ratio (M-H, Random, 95% CI)1.23 [0.48, 3.20]

 20 Amount of red blood cells administered3103Mean Difference (IV, Random, 95% CI)-111.74 [-594.13, 370.65]

    20.1 Trials with low bias risk
135Mean Difference (IV, Random, 95% CI)-600.0 [-899.18, -300.82]

    20.2 Trials with high bias risk
268Mean Difference (IV, Random, 95% CI)150.89 [-396.18, 697.96]

 21 Subjective overall quality of life assessment1897Mean Difference (IV, Fixed, 95% CI)-2.0 [-5.01, 1.01]

 22 Objective assessment of physical performance and dependency (Karnofsky)1897Mean Difference (IV, Fixed, 95% CI)-2.0 [-4.49, 0.49]

 23 Incidence of surgical intervention3103Risk Ratio (M-H, Fixed, 95% CI)1.04 [0.85, 1.27]

    23.1 Trials with low bias risk
263Risk Ratio (M-H, Fixed, 95% CI)0.93 [0.63, 1.36]

    23.2 Trials with high bias risk
140Risk Ratio (M-H, Fixed, 95% CI)1.12 [0.91, 1.38]

 24 Severity of sepsis I3156Mean Difference (IV, Random, 95% CI)-1.24 [-2.18, -0.29]

    24.1 Final MOF score among survivors, trials with low bias risk
188Mean Difference (IV, Random, 95% CI)-0.70 [-1.22, -0.18]

    24.2 Final MOF score among survivors, trials with high bias risk
268Mean Difference (IV, Random, 95% CI)-1.92 [-3.05, -0.78]

 25 Severity of sepsis II3102Mean Difference (IV, Fixed, 95% CI)-2.18 [-4.36, -0.00]

    25.1 Final APACHE I & II scores among survivors, trials with high bias risk
3102Mean Difference (IV, Fixed, 95% CI)-2.18 [-4.36, -0.00]

 26 Severity of sepsis III128Mean Difference (IV, Fixed, 95% CI)-0.2 [-0.66, 0.26]

    26.1 Final OSF score among survivors, trials with high bias risk
128Mean Difference (IV, Fixed, 95% CI)-0.2 [-0.66, 0.26]

 27 Severity of Ilness score149Mean Difference (IV, Fixed, 95% CI)-10.0 [-20.40, 0.40]

    27.1 Trials with low bias risk
149Mean Difference (IV, Fixed, 95% CI)-10.0 [-20.40, 0.40]

 28 Incidence of respiratory failure not present at admission62591Risk Ratio (M-H, Random, 95% CI)0.93 [0.76, 1.14]

    28.1 Trials with low bias risk
42524Risk Ratio (M-H, Random, 95% CI)0.96 [0.69, 1.32]

    28.2 Trials with high bias risk
267Risk Ratio (M-H, Random, 95% CI)0.83 [0.58, 1.20]

 29 Duration of mechanical ventilation3190Mean Difference (IV, Fixed, 95% CI)2.20 [-1.21, 5.60]

    29.1 Trials with low bias risk
1122Mean Difference (IV, Fixed, 95% CI)3.0 [-1.97, 7.97]

    29.2 Trials with high bias risk
268Mean Difference (IV, Fixed, 95% CI)1.49 [-3.18, 6.15]

 30 Length of stay in hospital3172Mean Difference (IV, Random, 95% CI)-1.86 [-11.38, 7.67]

    30.1 Trials with low bias risk
149Mean Difference (IV, Random, 95% CI)-10.80 [-20.49, -1.11]

    30.2 Trials with high bias risk
2123Mean Difference (IV, Random, 95% CI)2.97 [-4.01, 9.94]

 31 Mean length of stay in ICU6346Mean Difference (IV, Fixed, 95% CI)0.01 [-1.75, 1.76]

    31.1 Trials with low bias risk
1120Mean Difference (IV, Fixed, 95% CI)-1.70 [-5.09, 1.69]

    31.2 Trials with high bias risk
5226Mean Difference (IV, Fixed, 95% CI)0.63 [-1.42, 2.69]

 32 Overall mortality among septic population62601Risk Ratio (M-H, Fixed, 95% CI)0.95 [0.88, 1.03]

    32.1 Trials with low bias risk
32469Risk Ratio (M-H, Fixed, 95% CI)0.95 [0.88, 1.03]

    32.2 Trials with high bias risk
3132Risk Ratio (M-H, Fixed, 95% CI)0.90 [0.52, 1.57]

 
Table 1. Trials excluded (abstracts)


Trials excluded (abstracts)AT III mortality n/NControl mortality n/N

Balk 19954/172/17

Palareti 19956/309/29

Muntean 19899/458/53

Korninger 19870/50/5

Schuster 19979/2210/23

Paternoster 20000/180/21

 
Table 2. Refers to table of 'Comparisons and data' 01.01 to 01.06


Risk of bias analysisWeight
low-risk trials
RR (fixed) 95% CI, low risk of bias and
adequate quality
RR (fixed) 95% CI,
high risk of bias, inadequate or unclear

Overall mortality (quality)88.87%0.97 (0.88 to1.03)1.00 (0.78 to 1.27)

Overall mortality (random sequence generation)88.33%0.95 (0.88 to 1.02)0.99 (0.78 to 1.25)

Overall mortality (allocation concealment)90.22%0.96 (0.88 to 1.04)0.94 (0.72 to 1.21)

Overall mortality (blinding)90.67%0.95 (0.88 to 1.03)0.99 (0.76 to 1.28)

Overall mortality (completeness of follow up)88.85%0.97 (0.88 to 1.04)0.97 (0.77 to 1.23)

Overall mortality (Intention to treat)13.49%0.96 (0.77 to 1.19)0.96 (0.88 to 1.04)

Key: mortality n/N AT III 667/1708 (39.1%); n/N control 699/1750 (39.9%)RR (fixed) 0.96, 95% CI 0.89 to 1.03, Heterogeneity I² = 0 %, P = 0.81Test of interaction P = 0.17

 (a/A) PO2: ratio of arterial to alveolar oxygen pressure; ABG: arterial blood gas; AIDS: acquired immune deficiency syndrome; APACHE: acute physiology and chronic health evaluation; APTT: activated partial thromboplastin time; ARDS: acute respiratory distress syndrome; AT III: antithrombin III; BP: blood pressure
bid: bis in die, which means twice-daily; CNS: central nerve system; COPD: chronic obstructive pulmonary disease; CPP: cerebral perfusion pressure; CRP: C-reactive protein; CT: computer tomography; DIC: disseminated intravascular coagulation; FDP: fibriongen degradation products; FFP: fresh frozen plasma; FFS: fresh frozen serum; FHF: fulminant hepatic failure; g/l: gram per litre; GCS: Glasgow coma score; GI: gestosis index; h: hour; HELLP syndrome: haemolysis, elevated liver enzymes, low platelet count; HSCT: haematopoietic stem cell transplantation; ICP: intracranial pressure; ICU: Intensive care unit; ISS: injury severity score;ITT: Intention to treat; IU: international unit; IV: intravenous; IVH: intra ventricular haemorrhage; Kg: kilogram; LMWH: low molecular weight heparin; Min: minutes; ml: millilitre; mm Hg: millimetre mercury; MODS: multiorgan dysfunction syndrome; MOFS: multiorgan failure score; MPI: Mannheimer peritonitis index; n: number; NSAID: nonsteroid anti-inflammatory drugs; NYHA: New York Heart Association classification; OSFS: organ system failure score; PC: platelet concentration; PK: prothrombin complex; Plg: plasminogen; PPIC: plasmin-plasmin inhibitor complexes.
 
Table 3. Refers to tables of 'Comparisons and data': 01.07; 01.08; 01.12; 01.13; 01.14.


Subgroup analysisMortality
n/N AT III
Mortality
n/N Control
RR (fixed) (95% CI)

Follow up less than median of all trials87/319 (27.3%)92/360 (25.6%)0.96 (0.76 to 1.20)

Follow up greater than median of all trials508/1389 (36.6%)607/1390 (43.7%)0.97 (0.88 to 1.04)

Duration of Intervention less than/equal to 1 week647/1620 (39.9%)677/1630 (41.5%)0.96 (0.89 to 1.04)

Duration of Intervention greater than 1 week20/88 (25%)22/120 (18.3%)0.89 (0.59 to 1.34)

No adjuvant heparin *42/211 (19.9%)47/214 (22.0%)0.92 (0.67 to 1.27)

No adjuvant heparin §578/1368 (42.3%)608/1371 (44.3%)0.95 (0.88 to 1.04)

No adjuvant heparin §§200/563 (35.5%)229/560 (40.7%)0.89 (0.77 to 1.02)

Key: heterogeneity I2 = 0 %, P > 0.55*: Warren 2001 considered as a trial with adjuvant heparin therapy§: Warren 2001 considered as a trial without adjuvant heparin therapy§§: Data from Warren 2001 split based on heparin administration

 
Table 4. Refers to tables of ' Comparison and data' 01.09 to 01.11 and 01.32


Subgroup analysisMortality: AT IIIMortality: controlRR (fixed) (95% CI)

Trauma9/33 (27.3%)4/35 (11.4%)2.15 (0.81 to 5.72)

Obstetrics (maternal)0/880/87

Obstetrics (fetal & neonatal)5/78 (6.4%)5/80 (6.3%)1.03 (0.33 to 3.21)

Paediatrics *8/116 (6.9%)5/151 (3.3%1.49 (0.38 to 5.87)

Septic population590/1298 (45.5%)622/1303 (4.7%)0.95 (0.88 to 1.03)

Key: heterogeneity I2 = 0 %, P > 0.13*: RR (random) since I2 = 20.9 %, P = 0.26

 
Table 5. Refers to tables of ' Comparison and data' 01.15 to 01.19, 01.23, 01.28


Secondary outcomesAT III n/NControl n/NLow risk of bias trials RR (95% CI)High risk of bias trials RR (95% CI)Overall RR (95% CI)

Intracranial bleeding (1)36/122626/12281.62 (0.96 to 2.73)0.92 (0.51 to 1.66)1.26 (0.83 to 1.92)

Renal failure (2)2/333/33No trial0.71 (0.08 to 6.11)0.71 (0.08 to 6.11)

Pneumothorax (3)2/303/300.67 (0.12 to 3.71)No trial0.67 (0.12 to 3.71)

Complications other than bleeding (4)14/7533/1120.75 (0.18 to 3.07)0.72 (0.40 to 1.30)0.72 (0.42 to 1.25)

Bleeding events (5)312/1447199/14821.37 (1.04 to 1.82)1.23 (0.48 to 3.20)1.52 (1.30 to 1.78)

Incidence of surgical intervention (6)31/5130/520.93 (0.63 to 1.36)1.12 (0.91 to 1.38)1.04 (0.85 to 1.27)

Respiratory failure not present at admission (7)115/1293137/12980.96 (0.69 to 1.32)0.83 (0.58 to 1.20)0.93 (0.76 to 1.14)

Key: RR (fixed) in analysis 3, 4, 6 since I2 = 0%, P > 0.33RR (random) in analysis 1, since I2 = 9.2 %, P = 0.33RR (random) in analysis 2, since I2 = 28.2 %, P = 0.24RR (random) in analysis 5, since I2 = 0.3 %, P = 0.43RR (random) in analysis 7, since I2 = 31.6%, P = 0.22

 
Table 6. Refers to tables of ' Comparison and data' 01.20 to 01.22, 01.24 to 01.27, 01.2


Secondary outcomesAT III NControl NLow risk of bias trials MD (95% CI)High risk of bias trials MD (95% CI)Overall MD (95% CI)

Amount of red blood cell transfused (1)5053-600.00 (-899.18 to -300.82)150.89 (-396.18 to 697.96)-111.74 (-594.13 to 370.65)

Subjective overall quality of life assessment (2)460437-2.00 (-5.01 to 1.01)No trial-2.00 (-5.01 to 1.01)

Objective assessment of physical performance and dependency (3)460437-2.00 (-4.49 to 0.49)No trial-2.00 (-4.49 to 0.49)

Severity of sepsis (MOF score) (4)8076-0.70 (-1.22 to -0.18)-1.92 (-3.05 to -0.78)-1.24 (-2.18 to -0.29)

Severity of sepsis (APACHE I and II) (5)5250No trial-2.18 (-4.37 to 0.00)-2.18 (-4.37 to 0.00)

Severity of sepsis (OSF score) (6)1513No trial-0.20 (-0.66 to 0.26)-0.20 (-0.66 to 0.26)

Severity of illness score (7)2425-10.00 (-20.40 to 0.40)No trial-10.00 (-20.40 to 0.40)

Duration of mechanical ventilation (8)1081113.00 (-1.97 to 7.97)1.49 (-3.18 to 6.15)2.20 (-1.21 to 5.60)

Length of stay in hospital (9)8785-10.80 [-20.49 to -1.11]2.97 (-4.01 to 9.94)-1.86 (-11.38 to 7.67)

Mean length of stay in ICU (10)176170-1.70 [-5.07 to 1.69]0.63 (-1.42 to 2.69)0.01 (-1.75 to 1.76)

Key: Mean difference (MD) (fixed) in analysis 2, 3, 5, 6, 7, 8, 10, since I2 = 0%, P > 0.63MD (random) in analysis 1, since I2 = 87.7 %, P = 0.0003MD (random) in analysis 4, since I2 = 47.8 %, P = 0.15MD (random) in analysis 9, since I2 = 64.3%, P = 0.06