TRICS III

Published

Mazer CD, et al. "Restrictive or Liberal Red-Cell Transfusion for Cardiac Surgery". The New England Journal of Medicine. 2017. 377(22):2133-44.
PubMed • Full text • PDF • ClinicalTrials.gov

Clinical Question

In patients undergoing cardiac surgery who have a moderate-to-high risk of death, is a restrictive transfusion strategy applied throughout the perioperative period noninferior to a liberal transfusion strategy, in terms of major morbidities and mortality?

Bottom Line

A restrictive red-cell transfusion strategy (transfusing for hemoglobin <7.5 g/dL) resulted in fewer red-cell transfusions than a more liberal red-cell transfusion strategy (hemoglobin <9.5 g/dL) and was noninferior for major clinical outcomes among moderate- and high-risk patients undergoing on-pump cardiac surgery.

Major Points

Red-cell transfusions can be life-saving in many contexts, but unnecessary transfusions carry considerable morbidity and mortality. A growing body of evidence from randomized data supports more restrictive transfusion thresholds among patients in the ICU (TRICC), with sepsis (TRISS), and in acute upper GI hemorrhage. There is limited data supporting a particular transfusion threshold in the surgical setting. FOCUS, for example, demonstrated that among patients undergoing hip surgery, a liberal strategy provided no benefit over a more restrictive approach. Among patients undergoing cardiac surgery, the TITRe2 trial raised concerns because a restrictive transfusion strategy was associated with increased 90-day mortality compared to a liberal strategy. That result, coupled with limitation in the study's design, prompted additional studies.

The international Transfusion Requirements in Cardiac Surgery (TRICS) III trial enrolled over 5,000 adults with a moderate-to-high predicted risk of death undergoing on-pump cardiac surgery, including CABG and non-CABG procedures. Patients were randomized to a restrictive red-cell transfusion threshold in which a hemoglobin <7.5 g/dL triggered transfusion, or to a liberal threshold in which transfusion was administered for a hemoglobin <9.5 g/dL when in the ICU or <8.5 g/dL in a non-ICU hospital ward. Transfusions were administered intraoperatively and postoperatively based on hemoglobin measurements intraoperatively and through postoperative day 11 if the patient remained hospitalized. The trial's primary objective was to establish whether the restrictive transfusion strategy was noninferior to the liberal strategy with respect to the primary composite outcome of death from any cause, MI, stroke, or new renal failure requiring HD. The primary outcome occurred in 11.4% of patients in the restrictive group, compared to 12.5% of patients in the liberal transfusion group, with a risk difference that fell within the prespecified 3% margin for noninferiority (P<0.001 for noninferiority). Of note, in the restrictive group, 52.3% of patients received a transfusion after randomization, compared to 72.6% of patients in the liberal strategy group. The authors note limitations of the study which include lack of blinding (not feasible in a transfusion trial) and issues with generalizability as the trial enrolled only moderate- to high-risk patients. Ultimately the authors conclude that a restrictive strategy is feasible, safe, and leads to fewer units of blood transfused.

Guidelines

As of April 2018, no guidelines have been published that reflect the results of this trial.

Design

• International, open-label, randomized, controlled, noninferiority trial
• N=4,860 patients undergoing on-pump cardiac surgery
  • Restrictive strategy (hemoglobin <7.5 g/dL; n=2,430)
  • Liberal strategy (hemoglobin <8.5 or 9.5 g/dL; n=2430)
• Duration of follow-up: 28 days
• Analysis: Per-protocol
• Primary outcome: Composite outcome of death from any cause, MI, stroke, or new renal failure requiring HD

Population

Inclusion Criteria

• Age ≥18 years
• Planned cardiac surgery using cardiopulmonary bypass
• EuroSCORE of ≥6

Exclusion Criteria

• Unable to receive blood products
• Planned procedure is heart transplant or VAD placement
• Pregnancy or lactation

Baseline Characteristics

Comparisons are restrictive vs. liberal threshold groups

• Age: 72 vs. 72 years
• Male sex: 1553 (63.9%) vs. 1586 (65.3%)
• BMI: 28.1±6.0 vs. 28.0±5.2
• EuroSCORE: 7.9±1.8 vs. 7.8±1.9
• Previous cardiac surgery: 307 (12.6%) vs. 280 (11.5%)
• MI in previous 90 days: 562 (23.1%) vs. 601 (24.7%)
• LV function:
  • Good: 1485/2430 (61.1%) vs. 1523/2427 (62.8%)
  • Moderately reduced: 733/2430 (30.2%) vs. 710/2427 (29.3%)
  • Poor: 166/2430 (6.8%) vs. 156/2427 (6.4%)
  • Very poor: 46/2430 (1.9%) vs. 38/2427 (1.6%)
• Diabetes Mellitus: 646 (26.6%) vs. 686 (28.2%)
• Treated hypertension: 1797 (74.0%) vs. 1803 (74.2%)
• Emergency surgery: 37 (1.5%) vs. 34 (1.4%)
• Renal function:
  • Normal: 1090/2332 (46.7%) vs. 1071/2348 (45.6%)
  • Moderately impaired: 857/2332 (36.7%) vs. 866/2348 (36.9%)
  • Severely impaired: 355/2332 (15.2%) vs. 385/2348 (16.4%)
  • Use of dialysis: 30/2332 (1.3%) vs. 26/2348 (1.1%)
• Use of aspirin: 1274/2428 (52.5%) vs. 1293/2423 (53.4%)
• Hemoglobin (g/dL): 13.1±1.8 vs. 13.1±1.7
• Type of surgery:
  • CABG only: 622/2429 (25.6%) vs. 645/2430 (26.5%)
  • CABG and valve: 464/2429 (19.1%) vs. 472/2430 (19.4%)
  • CABG and other, nonvalve: 205/2429 (8.4%) vs. 203/2430 (8.4%)
  • Valve only: 703/2429 (28.9%) vs. 716/2430 (29.5%)
  • Other, non-CABG: 433/2429 (17.8%) vs. 394/2430 (16.2%)
• Duration of cardiopulmonary bypass (min): 120±59 vs. 121±57
• Intraoperative tranexamic acid: 2219/2428 (91.4%) vs. 2235/2428 (92.1%)

Interventions

• Participants were randomly assigned to the restrictive or liberal approach in a 1:1 ratio
  • Restrictive transfusion strategy: Transfuse RBCs if hemoglobin <7.5 g/dL
  • Liberal transfusion strategy: Transfuse RBCs if hemoglobin <9.5 g/dL in ICU or <8.5 g/dL in non-ICU hospital ward
• Hemoglobin measured preoperatively, intraoperatively, and days 1, 2, 3, 5, 7, 9, and 11 while hospitalized
• Study transfusions only given intra- and post-operatively
• Transfusion of 1 unit of RBCs was given if hemoglobin fell below assigned threshold, and was followed by reassessment of hemoglobin
• RBCs were given as soon as possible after a fall in hemoglobin below the assigned threshold, and were required to be administered within maximum time frames:
  • 2 hours intraoperatively
  • 18 hours in the ICU and step-down units
  • 40 hours on the non-ICU hospital ward

Outcomes

Comparisons are restrictive vs. liberal threshold groups.

Primary Outcomes

Death from any cause, MI, stroke, or renal failure requiring HD during hospitalization or within 28 days

11.4% vs. 12.5% (risk difference -1.11%; 95% CI -2.93-0.72; P<0.001 for noninferiority)

Secondary Outcomes

Death

3.0% vs. 3.6%

Stroke

1.9% vs. 2.0%

Myocardial Infarction

5.9% vs. 5.9%

New‐onset renal failure with dialysis

2.5% vs. 3.0%

Secondary Outcomes

Length of stay in ICU (median)

2.1 days vs. 1.9 days (HR 0.89)

Length of stay in hospital (median)

8.0 days in each group (HR 0.93)

Duration of mechanical ventilation (median)

0.38 days vs. 0.36 days (HR 0.94)

Prolonged low‐output state

40.9% vs. 40.6%

Infection

5.0% vs. 4.2%

Bowel infection

0.2% in each group

Acute kidney injury

34% vs. 33.9%

Seizure

2.1% vs. 1.7%

Delirium

12.6% vs. 10.9%

Encephalopathy

1.1% vs. 0.9%

Criticisms

• Blinding is not feasible in a transfusion study, which may introduce bias
• Low-risk patients were not included
• Does not address the possibility of lower thresholds being effective
• Does not demonstrate the efficacy of transfusion
• Cannot draw conclusions about noninferiority with regards to secondary outcomes
• The main analysis was per-protocol, which may have introduced bias

Funding

TRICS III was sponsored by grants from the: Canadian Institutes of Health Research, Canadian Blood Services, National health and Medical Research Council of Australia, and Health Research Council of New Zealand.