Low HDL levels in sepsis versus trauma patients in intensive care unit

Sepsis remains an important cause of mortality in intensive care unit (ICU) despite a better comprehension of its pathophysiology [1, 2]. Sepsis pathways are complex, including early activation of pro- and anti-inflammatory signaling, along with major non-immunological responses such as cardiovascular, neuronal, autonomic, hormonal and metabolic responses, as well as activation of coagulation [3].

High-density lipoproteins (HDLs) represent a family of particles characterized by their ability to transport cholesterol from peripheral tissues back to the liver that confers to them a cardiovascular protective effect [4, 5]. HDLs also appear to emerge as relevant players in both innate and adaptive immunity [6]. Because of their pleiotropic properties, including anti-inflammatory, anti-apoptotic or antioxidant functions, experimental studies have tested the efficacy of both reconstituted HDL and ApoA1 mimetic peptide perfusion in animal models of septic shock [7–9]. Some of these have demonstrated a protective effect of these HDL mimetics on mortality and shown a decrease in the inflammatory state [9–12]. Several clinical studies have been conducted to assess HDL concentration in septic conditions. Van Leeuwen et al. [13] underlined that in septic patients, HDL concentrations rapidly fall and can be reduced to 50% of recovery conditions. Chien et al. [14] have shown that low HDL levels at day 1 of severe sepsis were significantly associated with an increase in mortality and adverse clinical outcomes. In an observational study involving 151 consecutive septic patients, a low HDL concentration was independently related to 30-day mortality [15].

In trauma patients, a systemic inflammatory response syndrome (SIRS) results from several mechanisms such as neutrophil stimulation [16] or the release of endogenous factors termed damage-associated molecular patterns (DAMPs) after tissue injury [17] leading to pro- and anti-inflammatory cytokine over-expression [18].

Whereas HDL levels were shown to be decreased during sepsis, to our knowledge, no study has explored the lipid profiles during severe injury.

The aim of the present study was to compare the lipid profiles between trauma and septic patients in ICU and to characterize the association between lipid profile, severity and prognosis.

During sepsis, the dramatic decrease in HDL-C levels relative to normal conditions remains unexplained. Several hypotheses are possible, including an acute over-consumption of HDL particles, a decrease in liver HDL synthesis (especially in the case of hepatic failure) or an increased clearance following an upregulation of scavenger receptors such as SRB-1 [23]. HDL particles may also easily be redistributed from the intravascular to the extravascular compartment in the context of sepsis [13, 24]. One major action of HDL particles during sepsis is the clearance of bacterial components, such as lipopolysaccharides (LPS) [25]. After binding bacterial components, the clearance of HDL particles may also be increased in inflammatory conditions [25, 26]. Interestingly, in the population of septic patients studied here, HDL concentration was higher in patients with gram-positive versus gram-negative bacteria, suggesting that the clearance of HDL-LPS particles may be enhanced.

Both sepsis and trauma are characterized by an intense systemic inflammation that results from a disproportionate activation of inflammation signaling pathways [16, 27, 28]. In these two pathologies, cell injury initiates the secretion of different mediators such as DAMPs, which are recognized by the host, leading to a pro-inflammatory response [3, 29]. Also, similar leukocyte genomic responses to inflammation may be involved in both trauma and sepsis patients [30]. However, despite a common inflammatory background, we did not observe a similar lipid profile between sepsis and trauma patients. To our knowledge, except in spinal cord injury [31], lipid profile and especially HDL-C concentration have never been compared between trauma and septic patients. In this context, we show, for the first time, that in the early phase of trauma, HDL levels are preserved, contrasting with sepsis. The mechanisms that may explain the HDL-C level differences between trauma and sepsis remain unknown, but the interaction between HDL particles and LPS may be one of the causes [23]. The marked release of bacterial endotoxins may lead to an increased consumption of HDL particles.

This discrepancy between sepsis and trauma patients emphasizes the potential protective role of HDLs during sepsis. In view of HDL over-consumption and clearance in sepsis, their previously described pleiotropic effects may be lacking for the control of inflammation associated with this condition [7, 23].

Concerning the relationship between HDL-C levels and prognosis in septic patients, results of previous studies are controversial. On the one hand, Chien et al. [14] have shown that HDL-C levels in non-survivors were significantly lower than those of survivors from day 1 to day 4 and that HDL-C concentration at day 1 could predict the overall 30-day mortality rate. In their study, the cutoff value of 0.52 mmol/l had a sensitivity of 92% and a specificity of 80% for predicting the overall 30-day mortality rate. Barlage et al., in an observational study in ICU, also underlined the relationship between apo-A1 levels and 30-day mortality [15]. On the other hand, Lee et al. [32] reported no significant HDL-C differences between survivors and non-survivors in 117 septic patients. Only TG levels were associated with mortality.

Our study was not designed to test a potential difference in mortality but rather to compare trauma and septic patients. Comparing with the studies of Chien et al. and Barlage et al., the limited number of patients and the low proportion of sepsis in our cohort may explain this lack of relationship between HDL concentration and mortality. However, we found that death or a SOFA score >6 at day 3 was associated with low HDL-C levels. To our knowledge, only one previous study has explored the relationship between HDL concentrations at admission and organ failure progression: Cirstea et al. [33] have recently reported that low HDL-C levels at the time of admission for suspected sepsis were strongly and independently prognostic of subsequent multiple organ dysfunction.

One major finding of our study was that SOFA score, need for and duration of mechanical ventilation and length of ICU stay were all negatively associated with HDL-C levels. This result concerning length of ICU stay was similar to that found by Chien and al. [14]. Concerning SOFA score evaluation at day 1, our results are in line with those of Ferreira et al., showing that the sequential assessment of organ dysfunction during the first days of ICU admission was clinically relevant and a good indicator of prognosis [34].

Our study has some limitations. The first limitation is the marked difference in age between sepsis and trauma groups. A sensitivity analysis restricted to age-/sex-matched patients from the two groups was performed, and the lower levels in HDL-C in sepsis relative to trauma patients were confirmed. In addition, most cross-sectional studies in the general population report no difference in HDL levels between young and older subjects [35–38]. Taken together, age adjustment, sensitivity analysis and bibliographic data support our conclusion that HDL level differences are indeed due to septic conditions.

Second, we were not allowed to collect information on race and familial disorders of lipid metabolism which can affect the lipid levels.

Third, we have not measured inflammatory markers such as cytokines/chemokines in the two groups of patients. In this context, it is not possible to compare the inflammatory state between trauma and septic patients. However, SAPSII and SOFA scores were collected and no differences were found between trauma and septic patients, suggesting that these two groups of patients were comparable in terms of gravity.

Fourth, we analyzed lipid levels only at ICU admission. No additional measures were performed in the early phase of sepsis (i.e., at emergency room admission) and during the stay in ICU. Comparison between the basal lipid levels (i.e., without any episode of sepsis or trauma before admission to hospital) and the pathologic condition in the same patient would be of interest.

Fifth, peritonitis, urinary tract infection and angiocholitis are the main causes of sepsis recruitment in our surgical ICU, which is quite different from other ICUs and especially medical ICUs, which may explain some differences. The ability of HDL to bind and promote LPS clearance in cases of infection by gram-negative pathogens may also impact our results because our sepsis population involved more gram-negative pathogens.

Finally, additional studies with appropriate sample size calculation should be performed, with a larger number of patients. Our study is exploratory, with a limited sample size; our conclusion should therefore be taken with caution.

Regarding the main finding of our study, further experimental studies should be conducted to understand the reasons for the low HDL concentration during sepsis. In addition, a lipidomics approach may also be of interest. Dysfunctional HDLs during sepsis may represent an explanation for the observed organ dysfunction during sepsis [23]. The interaction between HDL particles and bacterial components, such as LPS, should also be addressed in future experimental studies. Finally, because of the pleiotropic effects of HDLs, therapeutic injection of HDL mimetic peptides or reconstituted HDLs may be of interest and should be tested in preclinical settings.

Our study has demonstrated that the lipid profile was different between sepsis and trauma ICU patients. HDL-C levels were low in septic patients, whereas their concentration was not altered in the case of trauma. This major difference reinforces the necessity to explore the therapeutic potential of HDL in sepsis.

In the subgroup of sepsis, SOFA score, need for and duration of mechanical ventilation and length of ICU stay were all negatively associated with HDL-C levels.

Our study suggests that HDLs may play an important role in sepsis. Further studies enrolling more patients, and especially septic patients, are necessary to test whether low HDL levels are a bystander prognosis marker of mortality or whether HDLs may limit the deleterious effects of sepsis.