ARTICLE
Philips CA, Maiwall R, Sharma MK, et al. Comparison of 5% human albumin and normal saline for fluid resuscitation in sepsis induced hypotension among patients with cirrhosis (FRISC study): a randomized controlled trial. Hepatol Int. 2021; May 25: doi 10.1007/s12072-021-10164-z. Online ahead of print.
Objective
To compare the outcomes of fluid resuscitation using normal saline vs 5% albumin in treating sepsis-induced hypotension in patients with cirrhosis
Background
Early fluid resuscitation has been proven to be a fundamental component in improving survival for patients with severe sepsis or septic shock. Crystalloid (most commonly normal saline) has traditionally been the fluid of choice for treating sepsis-induced hypotension. However, in patients with cirrhosis, various acute and chronic changes to their hemodynamics make treating sepsis-induced hypotension even more challenging. In previous studies, administration of albumin has been shown to increase intravascular volume by the amount infused, and in sepsis has been shown to stay within the intravascular compartment longer than normal saline. However, there is limited evidence to compare the choice of fluid and the protocols for fluid resuscitation in patients with cirrhosis and severe sepsis-related hypotension.
DESIGN
Prospective, open-labeled, randomized controlled trial in a single-center liver disease-focused hospital
POPULATION
Over the course of the study, 2462 patients were admitted to the intensive care and high dependency units, of which 1477 (59.9%) were diagnosed or suspected to have sepsis, and 438 (29.7%) had hypotension. A total of 308 patients met eligibility criteria and underwent randomization.
INCLUSION CRITERIA
- Patients with cirrhosis ≥ 18 years old
- Admitted to the ED, ICU or high dependency unit of the Institute of Liver and Biliary Sciences from February 2014 to June 2015
- Initial mean arterial pressure (MAP) <65 mm Hg
- Suspected (per history and examination) or documented evidence of infection
EXCLUSION CRITERIA
- Patients who had already received 500 mL of colloid or 2 L of crystalloid fluid within 12 hours of presentation
- Already on vasopressors/inotropes
- History of variceal bleeding
- Structural heart disease or severe cardiopulmonary disease
- Non-sepsis causes of hypotension
- Requiring emergent surgery
- Advanced hepatocellular carcinoma
- Previous adverse reaction to albumin
TREATMENT ARMS
- 5% human albumin IV bolus, 250 ml over 15–30 minutes, followed by a maintenance infusion of 50 ml/hour for a total of three hours (n = 154)
- 0.9% normal saline 30 mL/kg intravenous bolus over 15–30 minutes followed by a maintenance infusion of 100 ml/hour for a total of 3 hours (n = 154)
PRIMARY OUTCOME
- Improvement of hypotension, as measured by a MAP ≥ 65 mm Hg after 3 hours
SECONDARY OUTCOMES
- Improvement of hypotension, as measured by a MAP ≥ 65 mm Hg after 1 and 2 hours
- Comparison of heart rate, urine output and arterial lactate
- Overall mortality at 1 week
METHODS & LOGISTICS
MAP was measured via a left radial arterial line. Urine output was measured by placement of a urinary catheter, and renal perfusion was indirectly measured as hourly urine output. Cefepime was administered as per treating team discretion within one hour and modified as indicated afterwards. Patients who did not attain a MAP ≥ 65 mm of Hg at the end of the first hour were considered non-responders and were initiated on vasopressor support (norepinephrine +/- vasopressin) and treatment with fluids and management as per the primary care team. Patients who recovered from sepsis-induced hypotension in the time from 0 to 1 hour but developed hypotension (MAP < 65 mm of Hg) in the time period after the end of the 1st hour were also started on inotropes.
KEY RESULTS
|
|
Albumin group (n) |
Normal Saline group (n) |
p-value |
Odds Ratio (95% CI) |
|
MAP ≥ 65 mm Hg at 1 hour |
39 |
23 |
0.03 |
1.9 (1.08–3.42) |
|
MAP ≥ 65 mm Hg at 2 hours |
27 |
15 |
0.06 |
1.9 (1.00–3.87) |
|
MAP ≥ 65 mm Hg at 3 hours |
18 |
5 |
0.008 |
3.9 (1.42–10.9) |
Patients treated with 5% albumin showed better response with regard to improvement in MAP in the first hour and after 3 hours (the primary outcome), when compared to the normal saline group.
Significant improvement and steady maintenance of pulse without the development of tachycardia were better in patients treated with human albumin at different time periods.
|
|
Albumin group pulse (bpm) |
Normal Saline group pulse (bpm) |
p-value |
|
1st Hour |
99.5 ± 7.9 |
101.7 ± 8.8 |
0.02 |
|
2nd Hour |
97.9 ± 5.5 |
103.4 ± 6.7 |
< 0.001 |
|
3rd Hour |
96.6 ± 3.6 |
103.1 ± 5.9 |
< 0.001 |
Urine output was not significantly different between treatment arms. At the end of the 3rd hour, output was 36.9 ± 13.7 (ml/hr) in the albumin group vs 35.6 ± 20.1 (ml/hr) in the NS group, (p = 0.291).
Reduction in arterial lactate (mmol/L) was significantly greater in the albumin treated patients than saline, at the end of the second (6.2±2.1 vs 6.8±2.3, p<0.05) and third (5.6±2.2 vs 7.0 ± 2.1, p < 0.001) hours. The mean change in lactate between baseline and second hour, second and third hour and baseline and third hour in patients treated with albumin was 0.7 mmoL/L, 0.55 mmol/L, 1.25 mmol/L and in those treated with saline was 0.03 mmoL/L, − 0.22 mmol/L, − 0.19 mmol/L, respectively, which was statistically significant on repeated measures analysis (p < 0.001).
No deaths occurred in the first 3 hours of fluid therapy, and the albumin group had a higher mean survival (6.23 ± 0.16 days vs. 5.01 ± 0.22 days (p = 0.02) in the saline group).
PREDICTOR OF FLUID RESPONSE
On univariate analysis, female gender [odds ratio (OR), (95% confidence interval (CI), p value; 0.42 (0.22–0.80), p = 0.01], Child Pugh class [0.86 (0.74–0.99), 0.04], MELD [0.97 (0.94–1.00), 0.04] and SOFA [0.87 (0.77–0.98), 0.02] were found to be significantly associated with non-response to fluid resuscitation in the first hour.
On multivariate analysis, only female gender [2.3 (1.2–4.4), 0.02] and SOFA score [0.9 (0.8–1.0), 0.03] were found to be predictors of non-response to fluid resuscitation. A combination of female gender with SOFA cut off ≥ 11 (based on ROC, sensitivity 80%, low specificity of 20%; p = 0.015) correctly predicted non-responders in 73% of study patients.
ADVERSE EVENTS
All patients in both treatment groups tolerated fluid management per protocol, and there were no documented instances of new onset metabolic acidosis or pulmonary edema, or statistically significant difference in serum pH between groups after treatment.
STRENGTHS
- The first randomized controlled trial on fluid selection for resuscitation in patients with cirrhosis and sepsis-induced hypotension.
- Clear and objectively measured data points and outcomes (sustained hypotension, lactate, etc.) that have been proven to correlate with morbidity and mortality in sepsis.
LIMITATIONS
- Single-center study and lack of blinding
- Small sample size (n=154 in each treatment arm)
- Short time frame for mortality outcome (1 week)
- Precise values for shock are not known for this population, and targeted management goals may be different in cirrhotic patients.
- The low rate of patients reaching the endpoint (MAP ≥ 65 mmHg) at one hour could be due to the nature and degree of shock, or due to insufficient volume expansion.
- This study was targeted at the cirrhosis population as a whole and is unable to account for possible differences in outcome based on underlying etiology (alcohol, NASH, viral, etc.), which varies greatly in different populations.
- Further studies could explore whether larger volumes and/or longer infusion times of albumin could improve outcomes in patients with sepsis and cirrhosis.
- Patients with cirrhosis are commonly on beta-blockers, which may impact the measurement of MAP and heart rate.
EM TAKE-AWAYS
For patients with cirrhosis presenting with sepsis-induced hypotension, initial fluid resuscitation with 5% albumin produced a quicker and more sustained increase in MAP, short-term improvement in pulse and arterial lactate, and improved short-term mortality when compared to normal saline.