Hyper-oncotic albumin administration reduces mortality in acute Respiratory Distress Syndrome compared to crystalloid: a systematic review and meta-analysis.
Yan-Jie Zhang, Zhu Luo, Zhen Wang et al.
✦ AI-curated · Sources linked
In 30 seconds
This systematic review and meta-analysis assessed albumin versus crystalloid for volume replacement in adults with acute respiratory distress syndrome (ARDS), including three randomised trials and two non-randomised studies. Across the included studies, mortality was 33.2% (97/292) with albumin and 44.9% (133/296) with crystalloid (OR = 0.61, 95%CI 0.43-0.85, p = 0.004). However, randomised trials showed no significant mortality benefit, whereas the association was present in non-randomised studies. Hyper-oncotic albumin was associated with lower mortality only in non-randomised studies; the authors call for larger, high-quality trials.
Key findings
- Overall mortality was 33.2% (97/292) with albumin versus 44.9% (133/296) with crystalloid (OR = 0.61, 95%CI 0.43-0.85, p = 0.004).
- Randomised trials showed no mortality benefit (OR = 0.83, p = 0.54), while non-randomised studies found lower mortality with albumin (OR = 0.52, p = 0.002).
- Hyper-oncotic albumin was associated with lower mortality in non-randomised studies (OR = 0.40, p = 0.02), but not in randomised trials (OR = 0.74, p = 0.57).
- Oxygenation improved significantly on the first (p = 0.05) and second days (p < 0.0001); iso-oncotic albumin showed no mortality benefit (OR = 0.88, p = 0.72).
Why it matters
The pooled mortality association may appear clinically important, but it was not reproduced in the randomised-trial subgroup. The difference between randomised and non-randomised findings cautions against interpreting the overall estimate as evidence that albumin reduces mortality. The review also identifies early oxygenation changes, while highlighting that stronger trial evidence is needed to establish whether these findings translate into patient benefit.
What to keep in mind
The evidence base comprised only five publications, including two non-randomised studies; randomised trials did not show a mortality benefit. Trial sequential analysis indicated that further high-quality randomised trials are needed.