Objectives: Our objectives were to determine first whether albumin prevents heparin-binding protein-induced increased endothelial cell permeability and renal cell inflammation and second, whether a plasma heparin-binding protein-to-albumin ratio predicts risk of acute kidney injury, fluid balance, and plasma cytokine levels in septic shock.
Design: In vitro human endothelial and renal cell model and observation cohort of septic shock.
Settings: Research laboratory and multicenter clinical trial (Vasopressin and Septic Shock Trial).
Patients: Adult septic shock (norepinephrine dose > 5 μg/min for > 6 hr).
Interventions: In vitro: heparin-binding protein (or thrombin) was added with or without albumin to 1) human endothelial cell monolayers to assess permeability and 2) to human renal tubular epithelial cells to assess inflammation.
Measurements and main results: Transendothelial electrical resistance-a marker of permeability-of human endothelial cells was measured using a voltohmmeter. We measured plasma heparin-binding protein-to-albumin ratio and a panel of cytokines in septic shock patients (n = 330) to define an heparin-binding protein-to-albumin ratio that predicts risk of acute kidney injury. Albumin inhibited heparin-binding protein (and thrombin-induced) increased endothelial cell permeability at a threshold concentration of 20-30 g/L but increased renal tubular cell interleukin-6 release. Patients who developed or had worsened acute kidney injury had significantly higher heparin-binding protein-to-albumin ratio (1.6 vs 0.89; p < 0.001) and heparin-binding protein (38.2 vs 20.8 ng/mL; p < 0.001) than patients without acute kidney injury. The highest heparin-binding protein-to-albumin ratio (> 3.05), heparin-binding protein quartiles (> 69.8), and heparin-binding protein > 30 ng/mL were significantly associated with development or worsening of acute kidney injury (p < 0.001) in unadjusted and adjusted analyses and were robust to sensitivity analyses for death as a competing outcome. Heparin-binding protein and heparin-binding protein-to-albumin ratio were directly associated with positive fluid balance (p < 0.001) and with key inflammatory cytokines. Increasing quartiles of heparin-binding protein-to-albumin ratio and heparin-binding protein (but not albumin) were highly significantly associated with days alive and free of acute kidney injury and renal replacement therapy (p < 0.001), vasopressors (p < 0.001), ventilation (p < 0.001), and with 28-day mortality.
Conclusions: Albumin inhibits heparin-binding protein-induced increased human endothelial cell permeability and heparin-binding protein greater than 30 ng/mL and heparin-binding protein-to-albumin ratio greater than 3.01-but not serum albumin-identified patients at increased risk for acute kidney injury in septic shock.