FEAST

Published

Maitland K, et al. "Mortality after fluid bolus in African children with severe infection". The New England Journal of Medicine. 2011. 364(26):2483-2495.
PubMed • Full text • PDF

Clinical Question

In Sub-Saharan African pediatric patients presenting in shock, does early resuscitation with either saline or albumin compare to a no-bolus strategy in terms of all-cause mortality at 48 hours?

Bottom Line

Among Sub-Saharan African children meeting a particular definition of shock, saline and albumin resuscitation appears to increase mortality when compared to a no-bolus strategy.

Major Points

Despite a lack of clinical evidence supporting its efficacy, fluid resuscitation was incorporated into practice guidelines for the treatment of children with septic shock in 2002. Most patients receive crystalloids such as normal saline or lactated Ringer's solution (LR), although there are theoretical benefits to colloids such as albumin. While trials such as SAFE and ALBIOS attempt to address this debate, the question of whether fluid resuscitation is necessary at all has resurfaced. Resource-limited areas like sub-Saharan Africa often do not have adequate intensive care capabilities and often do not provide fluid resuscitation as their standard of care. This provided an ideal locale to compare the local standard of care (no fluids) with fluid resuscitation according to conventional Western guidelines.

Published in 2011, the Fluid Expansion as Supportive Therapy (FEAST) study enrolled 3,141 Sub-Saharan children with severe febrile illness and impaired organ perfusion, and randomized them to receive either albumin, saline, or no volume resuscitation. At 48 hours, mortality was higher with albumin (10.6%) and saline (10.5%) as compared to no volume resuscitation (7.3%). This was independent of malaria status, coma, severe anemia, base deficit, or lactate level.

The outcomes of this trial have been controversial. The trial design was limited by the lack of blood pressure as inclusion, and only one of three WHO criteria for shock was required. As such, not all children may have been in shock and may have received unnecessary fluid boluses. Additionally, half of the study population had malaria parasitemia; the study's results may therefore help address fluid resuscitation strategies in patients infected with malaria, but may not adequately address the fluid management in undifferentiated hypovolemic shock. Given these controversies, the WHO guidelines did not include the outcomes of FEAST in developing their guidelines on fluid resuscitation in children.

Before extrapolating the FEAST findings to the developed world, future research would need to explore similar fluid resuscitation strategies in the context of other resuscitative measures including readily available intensive care.

Guidelines

WHO Hospital Care for Children (2013, adapted)

• IV fluid administration for children in shock:
  • Without severe malnutrition:
    • Rapidly administer LR or normal saline 20 mL/kg
    • After first infusion, if no improvement repeat LR or normal saline 10-20 mL/kg rapidly; if bleeding, give blood 20 mL/kg over 30 minutes
    • After second infusion, if no improvement and signs of dehydration (including diarrhea), repeat LR or normal saline administration; consider epinephrine or dopamine if suspected septic shock, consider disease-specific management
    • If suspected malaria or anemia with shock, rapid fluid infusion should be administered with care; blood products should be given if severe anemia instead of LR or normal saline
  • With severe malnutrition:
    • Administer LR+D5, 1/2 strength Darrow's solution+D5, or NaCl 0.45%+D5 at 15 mL/kg over 1 hour
    • After first infusion if improved HR, increased pulse volume, or reduced respiratory rate and no evidence of pulmonary edema, repeat infusion at 15 mL/kg over 1 hour then switch to oral or NG rehydration with ReSoMal at 10 mL/kg/hr for up to 10 hours and re-initate feeding with starter F-75
    • If no improvement after two administrations of fluids at 15 mL/kg, give maintenance fluids at 4 mL/kg/hr and administer whole blood when available at 10mL/kg over 3 hours; use pRBC if cardiac failure, then initiate refeeding with starter F-75 an start IV antibiotics
    • If deterioration during IV rehydration (eg, rales, increase in HR by 15 BPM, or increased JVP), stop IV infusion

Design

• Multicenter, open-label, dual stratum, randomized controlled trial
• N=3,141
  • Albumin bolus (n=1,050)
  • Saline bolus (n=1,047)
  • No bolus (n=1,044)
• Setting: 6 centers in Sub-Saharan Africa (Kenya, Tanzania, Uganda)
• Enrollment: January 2009 to January 2011
• Mean follow-up: 4 weeks
• Analysis: Intention-to-treat
• Primary outcome: 48-hour mortality
• Secondary outcomes: Pulmonary edema, increased intracranial pressure, mortality, or neurologic sequelae at 4 weeks

Population

Inclusion Criteria

• Age 60 days to 12 years
• Severe febrile illness complicated by:
  • Impaired consciousness
  • Respiratory distress
  • Impaired perfusion
    • Capillary refill time >3 seconds
    • Low limb temperature gradient
    • Weak radial pulse
    • Severe tachycardia

Exclusion Criteria

• Severe malnutrition
• Gastroenteritis
• Non-infectious causes of shock (trauma, surgery, burns)
• Conditions which volume expansion was contraindicated

Baseline Characteristics

From the albumin group.

• Demographics: Mean age 23 months, female 45%
• Baseline health data:
  • Mid-upper arm circumference ≤11.5 cm: 2%
• Presentation data:
  • Temperature: >39C 23%, <36C 6%
  • Respiratory: Distress 83%, rate 58/min, O2 saturation <90% 25%
  • Cardiovascular: HR <80 bpm 1%, severe tachycardia 70%, weak radial pulse 20%, capillary refill time ≥2 seconds 68%, capillary refill time ≥3 seconds 25%

Interventions

• Randomized to bolus fluid resuscitation with either albumin or normal saline, or to the control group who received no fluid bolus
  • Stratum A (without severe hypotension)
    • Albumin bolus group: 20 ml/kg over 1 hour
    • Saline bolus group: 20 ml/kg over 1 hour
    • Control group: no-boluses given
  • Stratum B (severe hypotension)
    • Albumin bolus group: 40 ml/kg over 1 hour
    • Saline bolus group: 40 ml/kg over 1 hour
• Children received IV maintenance fluids (2.5-4 ml/kg/h); antibiotics; antimalarial, antipyretic, and anticonvulsant drugs; treatment for hypoglycemia and transfusion with 20 ml/kg of whole blood if Hgb <5 g/dl according to national guidelines.
• A structured clinical case-report was completed at admission, 1, 4, 8, 24, and 48 hours. Assessment of neurologic sequelae occurred at 4 weeks by an independent clinician unaware of study assignment.

Outcomes

Comparisons are albumin vs. saline vs. no bolus. RR is relative risk, CI is confidence interval.

Primary Outcome

All-cause mortality at 48h

10.6% vs 10.5% vs. 7.3%

Saline bolus vs. no bolus: RR 1.44; 95% CI 1.09-1.92; P=0.01

Albumin bolus vs. no bolus: RR 1.45; 95% CI 1.10-1.92; P=0.008

Albumin bolus vs. saline bolus: RR 1.00; 95% CI 0.78-1.29; P=0.96

Albumin or saline boluses vs. no bolus: RR 1.45; 95% CI 1.13-1.86; P=0.003

• Note significant mortality differences and associated relative risks at 48 hours, favoring the control group. Absolute risk increase of 3.3% (NNH=30)

Secondary Outcomes

At 48 hours

Pulmonary edema: 1.3% vs. 0.6% vs. 0.6%

Increased ICP: 1.5% vs. 1.7% vs. 1.1%

Severe hypotension: 0.1% vs. 0.2% vs. 0.3%

Allergic reaction: 0.3% vs. 0.4% vs. 0.2%

Pulmonary edema, increased ICP, or both: 2.6% vs. 2.2% vs. 1.6%

Saline bolus vs. no bolus: RR 1.34; 95% CI 0.72-2.51; P=0.34

Albumin bolus vs. no bolus: RR 1.57; 95% CI 0.87-2.88; P=0.10

Albumin bolus vs. saline bolus: RR 1.01; 95% CI 0.80-1.28; P=0.91

Albumin or saline boluses vs. no bolus: RR 1.46; 95% CI 0.85-2.53; P=0.17

All-cause mortality at 4 weeks

12.2% vs. 12.0% vs. 8.7%

Saline bolus vs. no bolus: RR 1.38; 95% CI 1.07-1.78; P=0.01

Albumin bolus vs. no bolus: RR 1.40; 95% CI 1.08-1.80; P=0.01

Albumin bolus vs. saline bolus: RR 1.01; 95% CI 0.80-1.28; P=0.91

Albumin or saline boluses vs. no bolus: RR 1.39; 95% CI 1.11-1.74; P=0.004

Neurological sequelae at 4 weeks

2.2% vs. 1.9% vs. 2.0%

Saline bolus vs. no bolus: RR 0.95; 95% CI 0.51-1.77; P=0.87

Albumin bolus vs. no bolus: RR 1.10; 95% CI 0.61-2.01; P=0.74

Albumin bolus vs. saline bolus: RR 1.16; 95% CI 0.63-2.14; P=0.62

Albumin or saline boluses vs. no bolus: RR 1.03; 95% CI 0.61-1.75; P=0.92

All-cause mortality or neurological sequelae at 4 weeks

15.2% vs. 14.6% vs. 11.1%

Saline bolus vs. no bolus: RR 1.31; 95% CI 1.04-1.65; P=0.02

Albumin bolus vs. no bolus: RR 1.36; 95% CI 1.08-1.71; P=0.008

Albumin bolus vs. saline bolus: RR 1.04; 95% CI 0.84-1.28; P=0.71

Albumin or saline boluses vs. no bolus: RR 1.33; 95% CI 1.09-1.64; P=0.005

Subgroup Analysis

For the primary outcome.

Malaria

Positive: RR 1.59; 95% CI 1.10-2.31

Negative: RR 1.43; 95% CI 1.01-2.04

Coma

Yes: RR 1.04; 95% CI 0.73-1.49

No: RR 1.69; 95% CI 1.21-2.36

Hemoglobin level

<5 g/dl: RR 1.71; 95% CI 1.16-2.51

≥5 g/dl: RR 1.31; 95% CI 0.93-1.84

Base deficit

≥8 mmol/l: RR 1.68; 95% CI 1.18-2.39

<8 mmol/l: RR 2.17; 95% CI 1.14-4.14

Randomization

Before amendment: RR 1.38; 95% CI 1.05-1.83

After amendment: RR 1.72; 95% CI 0.98-3.05

Other notable events

• In January 2011, the trial was stopped for futility, as the combined intervention arms had significantly higher mortality (RR 1.46, 95% CI 1.13-1.90, P=0.004) than the control arm.
• The risk of death was similar at 1 hour with a persistent trend for increased mortality up to day 2 in the two fluid bolus groups. Most (87%) of deaths occurred within the first 24 hours and very few occurred after 48 hours.
• In Stratum B (hypotensive patients) there was no difference between saline and albumin (RR 1.23; 95% CI 0.70-2.16; P=0.45).
• Investigators identified no subgroup (coma, malaria, severe anemia, age, gender, severe acidosis) for which fluid resuscitation was beneficial.

Criticisms

• As blood pressure was not a component of the inclusion criteria, it's possible that children who were not in shock were given volume resuscitation when it wasn't necessary.
• Only one of the three criteria of the World Health Organization definition of shock was required (delayed capillary refill, weak pulse, tachycardia), limiting the interpretation of shock. Moreover, only 2% of the children met WHO shock criteria
• While the authors excluded children with severe malnutrition by arm circumference, it is unclear how the authors further determined nutritional status.
• It is unclear what the cause of febrile illness was in the 43% of children without malaria
• The original manuscript did not analyze mortality by cause of death; however, the authors provided details assessing all-cause mortality in a subsequent analysis.
• Administration of large volumes of saline in children may cause cardiovascular collapse through induction of acidosis; 5% dextrose in water (D5W) can induce hyponatremia. Both of these fluids were used and may have resulted in worse outcomes than had LR or Hartmann's solution been used along with D5W or D10W
• The external validity of these findings in the developed world is uncertain, given access to ventilator support and intensive care in much of the developed world. Differences in access to resources, immunization and nutrition status of African children, as well as epidemiological differences in disease processes and prevalence (eg, malaria) further limits generalizability.

Funding

• Medical Research Council, United Kingdom
• Baxter Healthcare donated resuscitation fluids