Effects of intermittent hemodialysis on plasmatic levels of endocan

To the editor,

Endocan is a prognostic biomarker of pulmonary and systemic inflammatory states [1], such as acute respiratory distress syndrome or sepsis, with substantial proportion of subjects requiring RRT [2]. In previous correspondences, Honoré et al. raised the question of the actual impact of hemodialysis on plasmatic levels of endocan in critically ill subjects [3, 4].

To address this question, we retrospectively collected serial measurements of endocan performed for routine care in EDTA plasma from 11 patients undergoing intermittent hemodialysis in a 50-bed ICU in Lille, France. Hemodialysis was performed using either ELISIO (Nipro, France), a polyethersulfone membrane with very low adsorptive capacities or EVODIAL (Baxter, USA), a heparin-grafted AN 69 ST membrane with higher adsorptive properties. Patients’ characteristics on ICU admission and on day of hemodialysis were collected retrospectively.

We categorized measurements of plasma endocan according to timing of blood collection relatively to hemodialysis, as following: T0, within 1 h before start of hemodialysis; T1, 30 min to 60 min after start of hemodialysis; T2, 90 min to 120 min after start of hemodialysis, T3, 180 min to 240 min after start of hemodialysis; T4: 90 min to 120 min after end of hemodialysis.

We are reporting characteristics of patients and of hemodialysis in Table 1 and kinetics of plasma endocan during hemodialysis in Fig. 1. We find a significant variation of endocan during the time course of hemodialysis, with median [IQR] values measured at 10.9 [4.3; 14.9] ng/ml at T0, 15.4 [5; 19.4] ng/ml at T1, 14.9 [5.4; 18.8] ng/ml at T2, 12.4 [6.1; 20] ng/ml at T3, 10.1 [4.2; 16.1] ng/ml at T4 (p = 0.025). In addition, relatively to values observed at T0, we found significant increases of median [IQR] plasma endocan variations at T1 (+ 22% [+ 12%; + 72%], p = 0.006), T2 (+ 34% [-4%; + 68%], p = 0.019), and T3 (+ 27% [+ 5%; + 56%], p = 0.014), but not at T4 (+ 34% [-8%; + 57%], p = 0.053). Noteworthy, our analyses suggest that the variations of endocan depend on the type of membrane, with lower increases in the EVODIAL group (p = 0.009 by linear mixed model). This lower increase in endocan blood levels seems consistent with greater adsorption properties of EVODIAL membranes. However, because of the limited number of patients in the EVODIAL group, this result should be interpreted with caution.

Our results suggest an increase in levels of plasma endocan at the initial phase of hemodialysis. This may be explained by the response to vascular stress, resulting in an increase in secretion of endocan [1], or by hemoconcentration, which may occur during hemodialysis. Interestingly, the quick raise in endocan blood levels is consistent with previous data from a human LPS-induced endothelial stress model [5]. Additionally, our data tend to show a progressive decrease of plasma endocan during the time course of hemodialysis, possibly related to alleviation of endothelial stress or actual elimination of endocan. Adsorption on the membrane may explain this progressive decrease in blood levels of endocan, as suggested by the likely influence of adsorptive properties of the membranes on the variations of endocan. Conversely, the possibility that this decrease could be explained by diffusion mechanisms seems unlikely, given the 50 kDa molecular weight of endocan, exceeding the diffusion properties expected with our hemodialysis settings [6]. Similar mechanisms should be observed during continuous renal replacement therapy (CRRT), potentially resulting in greater removal of plasma endocan, because of longer RRT durations in CRRT.

Hence, these results suggest major interference of hemodialysis with blood concentrations of endocan, especially when highly adsorptive membranes are used, making it unreliable as a prognostic biomarker of pulmonary and systemic inflammation in critically ill patients undergoing hemodialysis.