Background
Methods
Data sources and search strategies
Study selection
Data extraction
Study endpoints
Assessment of methodological quality
Statistical analysis
Results
Search results
Study characteristics
TRICOP | TRISS | TRICC | |
---|---|---|---|
First author, year | Bergamin, 2017 [14] | Holst, 2014 [8] | Hebert, 1999 [15] |
Country | Brazil | Denmark, Sweden, Norway Finland | Canada |
No of site | 1 | 32 | 25 |
Inclusion period | 2012–2014 | 2011–2013 | 1994–1997 |
No of patients | 300 | 998 | 218 (subgroup) |
Type of patients | Cancer patients with septic shock (≧ 18 years old) | Patients with septic shock (≧ 18 years old) | Patients with severe infection or septic shock (≧ 16 years old) |
Source of infection [liberal (%) vs restrictive (%)] | |||
Lungs | 69 vs 61 | 52.2 vs 53.2 | |
Abdomen | 13 vs 17 | 39.9 vs 41.0 | Unknown |
Urinary tract | 9 vs 5 | 12.3 vs 11.6 | |
Other | 9 vs 17 | 21.4 vs 21.8 | |
Exclusion criteria (yes: excluded, no: not excluded) | |||
Life-threating/uncontrolled bleeding | Yes | Yes | Yes |
Acute myocardial ischemia | No | Yes | No |
Admission after cardiac surgery | No | No | Yes |
Hematologic cancer | Yes | No | No |
Acute burn injury | No | Yes | No |
End-stage renal disease | Yes | No | No |
Pregnancy | No | No | Yes |
Anticoagulation therapy | Yes | No | No |
Withdrawal from active therapy | Yes | Yes | Yes |
Hemoglobin threshold [liberal (g/dL) vs restrictive (g/dL)] | < 9 vs < 7 | < 9 vs < 7 | < 10 vs < 7 |
Severity of patients (SOFA score) [median (interquartile range)] | Liberal; 6 (5–9) | Liberal; 10 (8–12) | Unknown |
Restrictive; 7 (5–9) | Restrictive; 10 (8–12) | ||
Timing of randomization after ICU admission (h) [median (interquartile range)] | Within 6 h | Liberal; 20 (7–43) | Within 72 h |
Restrictive; 23 (7–50) | |||
Intervention period | ICU stay | ICU stay | ICU stay |
Leukodepletion | Yes | Yes | No |
Outcomes | Mortality (28, 60, 90 days) | Mortality (28, 60, 90 days) | Mortality (30 days) |
Ischemic events | Ischemic events | ||
Severe adverse reactions | Severe adverse reactions | ||
Use of life support at 28 days | Use of life support at 28 days | ||
Number of patients transfused in the ICU | Number of patients transfused in the ICU |
Outcome
Heterogeneity
Publication bias, risk of bias, and quality of evidence
Outcomes | Anticipated absolute effects* (95% CI) | Relative effect: OR (95% CI) | No. of participants (no. of studies) | Certainty of evidence (GRADE) | |
---|---|---|---|---|---|
Risk with restrictive RBC transfusion | Risk with liberal RBC transfusion | ||||
28- or 30-day mortality | 362 per 1000 | 360 per 1000 (276–454) | 0.99 (0.67–1.46) | 1516 (3 RCTs) | Lowa,b |
60-day mortality | 452 per 1000 | 429 per 1000 (312–554) | 0.91 (0.55–1.51) | 1298 (2 RCTs) | Lowb,c |
90-day mortality | 493 per 1000 | 453 per 1000 (323–588) | 0.85 (0.49–1.47) | 1298 (2 RCTs) | Lowb,d |
No. of patients transfused with RBC in the ICU | 585 per 1000 | 933 per 1000 (355–997) | 9.94 (0.39–250.88) | 1277 (2 RCTs) | Lowb,e |
Ventilation use at 28 days | 146 per 1000 | 147 per 1000 (81–250) | 1.01 (0.52–1.96) | 1277 (2 RCTs) | Lowb,f |
Vasopressor use at 28 days | 66 per 1000 | 69 per 1000 (28–161) | 1.06 (0.41–2.72) | 1277 (2 RCTs) | Lowb,g |
RRT at 28 days | 66 per 1000 | 64 per 1000 (41–99) | 0.97 (0.60–1.57) | 1277 (2 RCTs) | Moderateb |
No. of patients transfused with FFP in the ICU | 186 per 1000 | 205 per 1000 (163–257) | 1.13 (0.85–1.51) | 1277 (2 RCTs) | Moderateb |
No. of patients transfused with platelet in the ICU | 146 per 1000 | 165 per 1000 (120–221) | 1.16 (0.80–1.67) | 1277 (2 RCTs) | Moderateb |