Background and aims: Severe burn injury leads to a capillary leak syndrome, causing fluid and protein leak to the extravascular space, which in turn cause hypovolemia and burn shock. The administration of crystalloids in large amounts based on the Parkland formula or the modified Brooke formula (2-4 ml/kg/%) remains the standard. The early use of colloids such as albumin to decrease the fluid requirements, the mortality or the incidence of acute kidney injury has always been debated. We hypothesized that early hypoalbuminemia was an independent risk factor for mortality and that early albumin administration was associated with a decrease in mortality, acute kidney injury (AKI) and a decrease of fluid requirements.
Methods: This retrospective, single-center study was conducted in the intensive care burn unit of the Lausanne University Hospital (CHUV) between 01.01.2006 and 31.12.2018. Inclusion criteria were age ≥ 14 years and burns > 20 % TBSA. Exclusion criteria were admission > 8h following the burn accident, transfer during the first week to another burn unit or withdrawal of therapy during the first 72h. Data presented as median (IQR) or number (%). Logistic regression and multiple regression analysis were performed.
Results: Altogether 141 patients were included, with burns 35 (24-50) %TBSA, age 39 (26-56) years, 68.1% were male and 56.7% had inhalation injury. ABSI score was 8 (7-10). 17 (12%) patients died. Minimal albumin levels during the first 24 hours were significantly lower in non-survivors (15 (14-20) vs. 24 (19-31) g/l; p<0.001) and was found to be an independent risk factor for mortality when adjusted for ABSI scores (p < 0.001). AUC for first 24h minimal albuminemia was 0.81, and 22 g/l was identified as the best cut-off value to predict mortality (sensitivity 63.6 %; specificity 87.5 %). In univariate analysis, albumin 20% was administered more frequently to non-survivors versus survivors (12 (71%) vs. 30 (24%); p<0.001), and in increased quantities (47 (25-58) vs. 30 (10-40) grams; p=0.030). Albumin was given earlier in non-survivors with 9 (5-13) vs. 18 (14-20) hours; p<0.001). Acute kidney injuries were more frequent in patients that received albumin (29 (69%) vs. 33 (33%); p<0.001). After correction for the ABSI score, the logistic regression model showed that albumin administration increased the risk of developing AKI during the first 7 days (OR 1.03 (95% CI 1.00-1.05); p=0.035), without any significant reduction of fluid requirements. Increased quantities of albumin administered during the first 7 days was associated with a higher cumulated fluid balance, even after correction for the TBSA.
Conclusion: Hypoalbuminemia < 22 g/l during the first 24 hours was found to be an independent risk factor for mortality in severely burned patients. The administration of albumin was associated with a significant increase of AKI, without significant reduction of fluid requirement at 24h, and with an increase of fluid balance on day 7. The present study was underpowered to conclude regarding other endpoint such as mortality and dialysis requirements. In the absence of a prospective study clearly demonstrating of a benefit of albumin, its use should remain confined to extreme hypoalbuminemia.