Urinary albumin excretion is elevated in sepsis, but does not correlate with circulating VEGF-A levels

Abstract

Introduction: In critically ill patients admitted with SIRS, endothelial dysfunction leads to increased capillary permeability. In the glomerulus, it manifests as increased albumin excretion. Microalbuminuria is a common finding in various acute conditions like sepsis, trauma and surgery. Studies have shown increased levels of vascular endothelial growth factor-A (VEGF-A), a potent vascular permeability inducing agent, in LPS-induced endotoxemia, severe sepsis and septic shock. The pathogenic mechanism of the glomerular leakage of albumin in acute inflammatory conditions remains to be clarified. We wished to investigate the causative role of VEGF-A, in this regard which might have therapeutic implications. Methods: Prospective study in the 43-bed ICU of a tertiary care hospital. Of the consecutive patients admitted to the ICUs between September 2008 and January 2009, 597 patients were included, after excluding patients with ICU stay <24 hours, pregnancy, menstruation, anuria, hematuria and proteinuria due to renal and post- renal diseases. Of these, 30 consecutive patients with sepsis, severe sepsis and septic shock (sepsis group) and 30 randomly chosen patients without sepsis were recruited for the VEGF-A study. Spot urine samples for the albumin-creatinine ratio (ACR, mg/g) and serum for VEGF-A (ELISA) were collected on ICU admission. Correlation was analyzed using Spearman's correlation coefficient. Results: Sixty critically ill patients, with a median age (IQR) of 60 years (48 to 72), 39:21 male:female ratio, median APACHE II score (IQR) of 16 (7 to 23), and 3 median days of ICU stay, had a median ACR (IQR) of 125 mg/g (51.6 to 239.1) and a median VEGF-A (IQR) of 111 pg/ml (54.3 to 286.9) on ICU admission. The median ACR on admission in the sepsis group (n = 30) of 161.8 mg/g was significantly higher than the median ACR (78.3 mg/g) of the group without sepsis (n = 30) (P = 0.011). On statistical analysis, no significant correlation was obtained between levels of urinary ACR and serum VEGF-A (P = 0.327) in the entire group. Analysis for correlation between ACR and VEGF-A in the sepsis patients (P = 0.396) and in the group without sepsis (P = 0.518) yielded similar results. Conclusions: Despite a strong physiologic rationale, our pilot study did not show an association between microalbuminuria and VEGF-A in critically ill patients. Larger studies are needed to come to a definitive conclusion.