FOCUS

Published

Carson JL, et al. "Liberal or Restrictive Transfusion in High-Risk Patients after Hip Surgery". The New england Journal of Medicine. 2011. 365(26):2453-62.
PubMed • Full text • PDF

Clinical Question

Among high-risk patients undergoing hip surgery, does a liberal transfusion strategy (target hemoglobin >10 g/dL) improve survival or functional outcomes at 60 days compared to a restrictive strategy (target hemoglobin >8 g/dL)?

Bottom Line

Among high-risk patients undergoing hip surgery, a liberal transfusion strategy targeting a hemoglobin >10 g/dL does not improve survival or functional outcomes compared to a restrictive strategy targeting a hemoglobin of >8 g/dL.

Major Points

Red-cell transfusions are costly and potentially high-risk interventions in medical and surgical patients. Efforts have long been underway to limit unnecessary red-cell transfusions by implementing restrictive transfusion thresholds across patient populations. The role of a restrictive transfusion strategy was evaluated in several studies including TRICC in ICU patients, TRISS in patients with sepsis, and in Transfusion Strategies for Acute Upper Gastrointestinal Bleeding. Whether a restrictive transfusion strategy could be safely employed in patients undergoing elective hip surgery, especially those at high risk of a cardiovascular event, had not been demonstrated in a well designed randomized clinical trial.

The Transfusion Trigger Trial for Functional Outcomes in Cardiovascular Patients Undergoing Surgical Hip Fracture Repair (FOCUS) trial randomized 2,016 patients at high risk of a CV event to either a liberal transfusion strategy (transfusing RBCs to maintain hemoglobin >10 g/dL) or a restrictive transfusion strategy (transfusing RBCs to maintain hemoglobin >8 g/dL or for symptomatic anemia), following hip fracture surgery. The primary outcome was death or inability to walk 10 feet without human assistance at 60 days. The study population's mean age was approximately 80 years, and 75% patients were men, reflective of the demographics of patients with hip fracture. There was a 1.3-g/dL difference in pre-transfusion hemoglobin between the liberal and restrictive groups, and the median number of red-cell transfusions was 2 in the liberal group and 0 in the restrictive transfusion group. At 60 days, the rate of death or inability to walk unassisted was similar between groups (35.2% vs. 34.7%; P=0.90).

The FOCUS trial was limited by its mid-protocol expansion of eligibility criteria, potential lack of power to demonstrate a between-group difference in the primary outcome, and use of a composite outcome in which individual components may not be equal. Nevertheless the results of FOCUS are largely in line with other related trials to date, including TRICC and TRISS, and adds to the growing body of literature supporting the use of a restrictive transfusion strategy in most patients.

Guidelines

AABB Red Blood Cell Transfusion Thresholds and Storage (2016, adapted)

• For patients undergoing orthopedic surgery or cardiac surgery and those with preexisting cardiovascular disease, the AABB recommends a restrictive RBC transfusion threshold (hemoglobin level of 8 g/dL; strong recommendation, moderate quality evidence).
• The restrictive hemoglobin transfusion threshold of 7 g/dL is likely comparable with 8 g/dL, but RCT evidence is not available for all patient categories. These recommendations apply to all but the following conditions for which the evidence is insufficient for any recommendation: acute coronary syndrome, severe thrombocytopenia (patients treated for hematological or oncological disorders who at risk of bleeding), and chronic transfusion–dependent anemia.

Design

• N=2,016 high-risk adults undergoing hip surgery with hemoglobin <10 g/dL
  • Liberal strategy (n=1,007)
  • Restrictive strategy (n=1,009)
• Setting: 47 clinical sites in the US and Canada
• Enrollment: 2004-2009
• Follow-up: 60 days
• Analysis: Mantel-Haenszel method for primary analysis
• Primary outcome: Death or inability to walk without assistance at 60-day follow-up
• Secondary outcomes: In-hospital MI, unstable angina, or death for any reason (assessed individually and as a combined outcome)

Population

Inclusion Criteria

• Age ≥50 years
• Undergoing primary surgical repair of a hip fracture
• Hemoglobin level <10 g/dL within 3 days after surgery
• Established or at high risk of CV disease:
  • Original protocol required established cardiovascular disease defined as a history of ischemic CAD, EKG evidence of prior MI, history of or active HF or PVD, or history of stroke/TIA.
  • Expanded protocol in 2005 allowed enrollment of patients at high-risk of CV disease, defined as the presence of HTN, DM, HL, current tobacco use, or creatinine >2 mg/dL.

Exclusion Criteria

• Unable to walk without human assistance before hip fracture
• Declined blood transfusions
• Multiple trauma (planned or underwent surgery for non-hip related traumatic injury)
• Pathologic hip fracture associated with malignancy
• Acute MI within 30 days prior to randomization
• Previously participated in the trial with contralateral hip fracture
• Symptoms associated with anemia including ischemic chest pain
• Actively bleeding at time of potential randomization

Baseline Characteristics

From the liberal strategy group.

• Mean age: 82 years
• Female: 75.2%
• Ethnicity: White 93.7%, Black 4.0%, Asian 1.4%, Other 0.9%
• Residence in the United States 60.5%
• Cardiovascular disease
  • Any 63.3%
  • Coronary artery disease 39.9%
  • Congestive heart failure 18.3%
  • Cerebrovascular disease 24.7%
  • Peripheral vascular disease 11.6%
• Cardiovascular risk factors
  • Hypertension 82.2%
  • Diabetes mellitus 25.1%
  • Hypercholesterolemia 34.6%
  • Tobacco use 11.6%
  • Creatinine >2.0 mg/dL 8.3%
  • Chronic lung disease 18.8%
  • History of dementia or confusion 30.8%
  • History of cancer 18.0%
• Type of hip fracture:
  • Femoral neck 43.0%
  • Intertrochanteric 51.0%
  • Subtrochanteric 8.8%
  • Reverse oblique 1.3%
• Type of anesthesia
  • General 54.0%
  • Spinal 45.5%
  • Other 0.5%
• American Society of Anesthesiology Risk Factor Score 3.0±0.6
• Residence:
  • Home or retirement home 88.8%
  • Nursing home 10.3%
  • Other 0.9%

Interventions

• Patients were randomized equally to one of two groups:
  • Liberal strategy, in which 1 unit of RBCs was administered to maintain hemoglobin >10 g/dL; the post transfusion hemoglobin was evaluated and if it remained <10 g/dL then an additional unit of RBCs was transfused
  • Restrictive strategy, in which transfusions were administered if symptoms or signs of anemia developed (eg, cardiac chest pain, HF, or unexplained hemodynamic instability) or at physician's discretion for a hemoglobin of < 8 g/dL

• Hemoglobin monitoring was performed on hospital days 1, 2, 4, and 7 after randomization, and thereafter as clinically indicated.

• Investigators were to follow the transfusion strategy until patient discharge or 30 days post randomization, whichever came first.

Outcomes

Comparisons are liberal vs. restrictive transfusion strategy.

Primary Outcomes

Death or inability to walk 10 feet without assistance at 60 day follow-up

35.2% vs. 34.7% (OR 1.01; 95% CI 0.84-1.22; P=0.90)

Secondary Outcomes

Death at 60 days

7.6% vs. 6.6% (OR 1.17; 99% CI 0.75-1.83)

Inability to walk 10 feet without assistance at 60 days

27.6% vs. 28.1%

MI, unstable angina, or in-hospital death

4.3% vs. 5.2% (OR 0.82; 99% CI 0.48-1.42)

MI

2.3% vs. 3.8% (OR 0.60, 99% CI 0.30-1.19)

Unstable angina

0.2% vs. 0.3% (OR 0.67, 99% CI 0.06-7.03)

In-hospital death

2.0% vs 1.4% (OR 1.44, 99% CI 0.58-3.56)

Number of RBC units transfused (median)

2 vs. 0

Lower extremity physical ADLs

5.1 vs. 4.3 (P=0.85)

Instrumental ADLs

3.7 vs. 3.7 (P=0.94)

FACIT-Fatigue scale

41.8 vs. 42.3 (P=0.26)

Subgroup Analysis

Subgroup analysis of the primary outcome demonstrated a difference based on patient sex. In the liberal vs. restrictive transfusion strategy groups, the OR for having a primary outcome event was 1.45 in men versus 0.91 in women (P=0.03). There were no significant interactions with age, race, and cardiovascular disease status

Adverse Events

Isolated troponin elevation

6.2% vs. 5.9% (OR 1.06; 99% CI 0.65-1.71)

HF

2.7% vs. 3.5% (OR 0.77; 99% CI 0.39-1.50)

Stroke or TIA

0.8% vs. 0.3% (OR 2.69; 99% CI 0.47-15.42)

New infiltration on CXR

6.0% vs. 4.8% (OR 1.27; 99% CI 0.76-2.12)

Wound infection

1.4% vs. 0.8% (OR 1.76; 99% CI 0.56-5.56)

DVT or PE

1.2% vs. 0.8% (OR 1.51; 99% CI 0.46-4.92)

Transfer to ICU

3.0% vs. 2.9% (OR 1.04; 99% CI 0.53-2.05%)

Criticisms

• Inclusion criteria were altered mid-study from patients with proven cardiovascular disease to patients with risk factors including hypertension, diabetes, and hypercholesterolemia, such that at the end of enrollment only about 60% of patients had proven CV disease. While the authors report "no important treatment interaction with cardiovascular disease status," there is no presented subgroup analysis. In terms of drawing conclusions for patients with CV disease, this study is inherently underpowered and extrapolated from patients with risk factors.
• The paper states they required 2600 patients for 90% power at an alpha level of 0.05 to detect a 7 percentage point difference in the primary outcome, yet they only enrolled 2016, of which 748 were enrolled based on risk factors. The authors state that the 95% confidence interval upper limit for primary outcome is a 3.7% increase with the restrictive transfusion policy. Thus the study may have not had sufficient power to detect the prespecified difference in the primary outcome.
• The study did not analyze patients by cause of their anemia; thus patients with iron deficiency, anemia of chronic disease, and other causes of anemia were considered as a single group. Since the management of anemia differs according to the etiology of the anemia (eg, transfusion may be harmful in patients with anemia of chronic disease), the study's conclusions may not apply equally to patients with different causes of anemia.
• The use of a composite endpoint of death or inability to walk independently assumes relatively equal weight for each of these individual endpoints, which is not the case.

Funding

Funded by the NHLBI. Dr. Carson also reports receiving grant support to his institution from Amgen.