The relationship between hemorheological parameters and mortality in critically ill patients with and without sepsis

Abstract

PURPOSE: The prognostic scoring systems for mortality of intensive care patients estimate clinical outcome using several physiological and biochemical parameters. In altered hemodynamic conditions of critically ill patients, hemorheological variables may play a significant role in appropriate tissue perfusion. We investigated if hemorheological parameters are altered in critical status and if they could be markers of mortality.

METHODS: 112 patients (67.8 ± 12 years, 58 males, 54 females) treated in intensive care unit with different non-surgical diseases were investigated. Routine laboratory parameters and prognostic scores were determined and hemorheological variables (hematocrit, plasma and whole blood viscosity, red blood cell aggregation and deformability) were measured on the 1st and the 2nd day after admission.

RESULTS: ICU scores predicted 35.2–41.3% mortality rate, real mortality in intensive care unit was 37.5%, while 30-day mortality was 46.6%. Whole blood viscosity (WBV) and red blood cell (RBC) deformability were lower, red blood cell aggregation was higher in septic than in nonseptic patients (p < 0.05). In septic patients calcium was increased, osmolality was decreased, while in nonseptic patients WBV and RBC aggregation were higher in nonsurvivors compared to survivors (p < 0.05). Worsening of RBC deformability from day 1 to day 2 predicted higher mortality (p < 0.05).

CONCLUSION: Calcium and osmolality level were associated with outcome in sepsis. Whole blood viscosity, red blood cell aggregation and change in red blood cell deformability could predict mortality in nonseptic patients and they may add prognostic information over the ICU scores. Further investigations are needed to evaluate the benefit of our findings in clinical practice.

Keywords

Critically ill patients sepsis viscosity red blood cell aggregation red blood cell deformability

1 Introduction

Microcirculation has a crucial role in oxygen delivery and maintenance of tissue perfusion. It may be a reason why multiple organ failure (MOF) can develop in spite of the correction of global hemodynamic parameters. Sepsis is characterised by profoundly disturbed microcirculation with the decrease in the density of capillaries, increase in non-perfused and intermittently perfused vessels and functional shunting. Prior researches have indicated that these alterations can have a prognostic value in septic shock [7 , 27]. Hemorheological properties, which are important factors of microcirculation, can be essential in critical conditions, especially in sepsis.

Previous publications investigated hemorheological parameters mainly in sepsis. They suggested that red blood cell deformability is reduced in sepsis and it can be a marker of the severity of sepsis, furthermore it can refer to impaired oxygen utilization and multiple organ damage [5 , 28–32]. Red blood cell aggregation is increased in sepsis and it correlates with prognostic scoring systems [2 , 28]. Macrorheological factors, like whole blood viscosity (WBV) and plasma viscosity (PV) can be altered in septic and also in nonseptic patients [2]. Only one recent study researched these variables in correlation with outcome and highlighted the potential effect of microrheology [9].

Estimating mortality risk in intensive care unit is complicated although several mortality scoring systems have been developed in the past decades. The conventional systems (Acute Physiology and Chronic Health Evaluation (Apache) II and Simplified Acute Physiology Score (SAPS) II) use basic hemodynamic, laboratory (electrolytes, kidney function tests, arterial blood gases, white blood cell count), Glasgow Coma Scale and some anamnestic data, while in novel scoring systems the admission diagnoses and medical history are more pronounced. In spite of these steps of evolution only few new laboratory parameters were added (glucose and albumin to Apache IV and platelet count to SAPS III), but except for hematocrit in Apache systems there are no hemorheological markers in any of the scoring systems [16 , 36].

Nevertheless, the role of hemorheological parameters among critical conditions in a heterogeneous intensive care population and the possibility of them being prognostic markers remained unclear. This report describes the relationship among hemorheological parameters, mortality and clinical outcome in a heterogeneous population in an Intensive Care Unit.

2 Methods

2.1 Patients and study design

112 patients treated in intensive care unit (ICU) with different non-surgical diseases were recruited. Exclusion criteria were age under 18 years, admission from another ICU, and ICU readmission. Blood samples were drawn from patients within the first 24 hours after ICU admission. Blood sampling for hemorheological parameters was performed on the 2nd day in 83 patients as well (others deceased or were discharged to other non-intensive care department), and the change of values (Δ) was calculated. At ICU admission the presence of sepsis (identified using standard criteria as the presence of infection together with evidence of a systemic inflammatory response syndrome) and diagnosis were recorded, and ICU mortality scores (Acute Physiology and Chronic Health Evaluation (APACHE) II and IV score, Simplified Acute Physiology Score (SAPS) II and III [16 , 36]) were calculated. Hemodynamic parameters and received therapy in ICU were assessed. There was no significant difference between septic and nonseptic patients regarding age, gender, heart failure and respiratory failure, but acute renal failure and hematological disorders were more common in the septic group. In therapy mechanical ventilation (41.3% in nonseptic, 60.6% in septic), vasopressor (52.2% nonseptic, 81.8% in septic), and antibiotics (28.3% in nonseptic, 75.8% in septic) were more frequent in septic patients.

Mortality was followed up to 30 days.ICU mortality was 37.5%, while 30-day mortality was 46.6%. Demographic characteristics, ICU score values and diagnosis are showed in Table 1.

2.2 Laboratory measurements

Blood samples were taken from the patient’s arterial catheter. Routine laboratory (electrolytes, osmolality, renal function tests, liver function tests, coagulation markers, fibrinogen, total protein, albumin, C-Reactive Protein, procalcitonin, arterial blood gases, and complete blood count) and hemorheological parameters were determined.

2.3 Hemorheological measurements

Arterial blood was collected into K₂EDTA (10.8 mg in 6 ml) coated Vacutainer tubes. Hemorheological measurements were carried out within 2 hours after blood sampling [22, 36].

Hematocrit (Hct) was measured in a microhematocrit centrifuge (Haemofuge, Heraeus Inst., Germany) by using native capillaries at room temperature (22 ± 1°C). Plasma fibrinogen concentration was determined from sodium citrate anticoagulated blood by Clauss method [18]. The other hemorheological measurements were performed at 37°C.

Whole blood viscosity (WBV) was determined with Brookfield DV-III Ultra LV cone-plate rotational viscometer (Brookfield Engineering Laboratories Inc, Middleboro, USA) at 15 shear rates (from 400 to 50 s^–1) [1]. Plasma viscosity (PV) was measured with Hevimet 40 capillary viscometer (Hemorex Ltd., Budapest, Hungary) [20]. Plasma was prepared by centrifuging whole blood samples for 10 minutes at 1500 g.

Red blood cell aggregation was measured with LORCA aggregometer (Laser-assisted Optical Rotational Cell Analyzer; R&R Mechatronics, Hoorn, The Netherlands) using laser light back-scattering syllectometry. Blood samples were preoxygenated. Aggregation Index (AI, denoting to the proportion of aggregation – higher value means larger aggregation) and t½ (denoting to the swiftness of aggregation – lower value means faster aggregation) and RBC disaggregation threshold shear rate (γ, the smallest shear rate required for the complete disaggregation of erythrocytes – higher value means stronger aggregation) were calculated [10].

Red blood cell deformability was measured with LORCA ektacytometer by diffraction ellipsometry technique and Elongation Index (EI, 0 referring to no deformation, 1 to infinite deformation) was determined at 9 different shear stresses (from 0.3 to 30 Pa). A polyvinylpyrrolidone solution (32.6 mPas viscosity) was used to suspend blood sample [11].

2.4 Ethical aspects

The study was approved by the Regional Ethics Committee of the University of Pecs and an informed consent was signed by all subjects or their relatives (number of approval: 5828).

2.5 Statistical analysis

IBM SPSS statistical software version 22 was used. Data are expressed as means ± SD. Variables that are considered normally distributed with Shapiro-Wilk test were evaluated by Independent sample T-test. Nonparametric Mann-Whitney U-test was applied for non-normally distributed variables. For survival analysis Kaplan Meier test and Cox proportional hazard model was used. Significance level was defined as p < 0.05.

3 Results

Table 2 shows hemodynamic and laboratory parameters of survivors and nonsurvivors. Heart rate was higher, blood pressure was lower in nonsurvivors. Calcium (Ca) level was decreased in nonsurvivors and osmolality was increased in patients who did not survive 30 days, and in septic patients. Kaplan-Meier analysis and Cox proportional hazard model represented increased changes in Ca level with about twice higher mortality risk (Hazard ratio = 2.587, CI: 1.045–6.408) and patient with Δosmolality above the median have about 3-times lower survival rate (HR = 2.986, CI: 1.306–6.827, Fig. 1). Total protein and albumin level were lower in nonsurvivors. INR was increased in patients who died in the ICU. Inflammatory parameters were detected higher and nonsurvivors were more prone to metabolic acidosis and elevation of lactate level. Fibrinogen was not different between survivors and nonsurvivors, but its 1st–day level was higher in septic than in nonseptic patients, while the change between the measurements was higher in nonseptic patients.

Table 3 describes hemorheological parameters. There was no difference in capillary hematocrit level between survivors and nonsurvivors. The 1st–day measurements showed no differences among the groups, but the 2nd-day WBV was higher in nonsurvivors among nonseptic patients. No differences could be detected in PV values, nevertheless an increasing tendency from the 1st to the 2nd measurement was found in nonsurvivors. In nonseptic patients Kaplan-Meier analysis showed that patients with WBV values above the median had poorer survival (Fig. 1), with about 4-times higher mortality risk according to Cox proportional hazard model (HR of 1st–day WBV at 90 s^–1 = 3.968, CI: 1.090–14.453).

Although no difference was found in the total population, among nonseptic patients nonsurvivors showed increased RBC aggregation in the first day (Table 3). Kaplan-Meier analysis showed a significant difference in 30-day survival of patients with aggregation above or below the median in the total population (t½ HR = 0.570 CI: 0.325–1.000). It was more prominent in the nonseptic group (Fig. 1) with about 4-times higher Hazard ratio for higher aggregation (HR = 4.413, CI: 1.228–15.858), although in the septic group no difference was found.

In nonsurvivors RBC deformability (at higher shear stresses) showed worsening from the 1st to the 2nd day, and it was found to be lower on the second day compared to survivors (EI at 30-3 Pa, Table 3) in the whole examined population. Although survival analysis referred to lower survival in patients whose RBC deformability worsened from the 1st to the 2nd day compared to those whose RBC deformability improved (HR of ΔEI at 30 Pa = 2.669, CI: 1.375–5.181), no relationship was found in sepsis during subgroup analysis. In nonseptic patients survival analysis represented 7-times higher mortality risk in patients with worsened RBC deformability (Fig. 1, HR of ΔEI at 30 Pa = 7.647, CI: 1.617–36.173).

1st–day fibrinogen was higher in sepsis and it decreased more than in nonseptic patients. Although the 1st day WBV and RBC aggregation were not different between septic and nonseptic patients, changes of WBV between the 1st and 2nd day decreased in sepsis, and changes of RBC aggregation increased in nonseptic patients. RBC deformability (EI at 30–1.69 Pa) was impaired in septic patients both on the 1st and on the 2nd day (Table 4).

Red blood cell aggregation correlated significantly with prognostic scores (1st–day AI, r_Apache IV = 0.213; r_{Apache IV LOS} = 0.277; r_SAPS II = 0.185; r_SAPS III = 0.187). Correlation with red blood cell deformability was significant with 2nd-day EI at 30 Pa (r_Apache II = –0.247; r_SAPS III = –0.312), as well as with ΔEI at 30 Pa (r_Apache IV = –0.285, r_{Apache IV LOS} = –0.282, r_SAPS II = –0.226, r_SAPS III = –0.215). There was no relationship between ICU scores and macrorheological parameters, but fibrinogen level correlated significantly with scores (r_Apache II = –0.213, r_Apache IV = –0.224, r_{Apache IV LOS} = –0.329, r_SAPS II = –0.206).

To evaluate if hemorheological parameters could provide further information about mortality risk to ICU scores, dichotomised hemorheological parameters (AI, WBV- being lower or higher than the median; ΔEI at 30Pa- positive or negative) as categorical variables were added to ICU scores in Cox proportional hazard models. In septic patients none of these parameters remained significant. In nonseptic patients higher AI, higher WBV and negative ΔEI meant increased mortality risk in the various models (Hazard ratios [CI]: AI_APACHE II = 4.151 [1.144–15.066], WBV_APACHE II = 5.047 [1.247–18.912], ΔEI_APACHE II = 10.329 [2.020–52.829]; WBV_APACHE IV = 3.981 [1.070–14.807], ΔEI _APACHE IV = 6.986 [1.457–33.487]; ΔEI _SAPS II = 7.017 [1.477–33.334]; ΔEI _SAPS III = 6.060 [1.257–29.209]).

4 Discussion

Several researches investigated hemorheological parameters in critically ill patients in the past few decades. Deteriorated red blood cell deformability in sepsis was described in animal models [5 , 30] and also in patients [2 , 32]. These studies suggested that decreased deformability can be a consequence of reactive oxygen superoxide, 2.3 diphosphoglycerate increase or the presence of lipopolysaccharide, and it can be a marker of the severity of sepsis. Increased RBC aggregation in sepsis [2 , 32] and altered macrorheological parameters [2 , 25] in critically ill patients are also known. However, only a few studies can be found in correlation with survival [2, 9] or ICU scoring systems [2 , 32].

In routine laboratory examination we found inflammatory parameters higher, albumin and total protein lower in nonsurvivors than in survivors, and the deterioration of ABG parameters were more characteristic in nonsurvivors. Interestingly, sodium, potassium, glucose, renal function parameters, bilirubin, and Hct were not different in spite of the fact that they are main parts of ICU scores. Other surprising results were the lower Ca, INR and the higher osmolality in nonsurvivors that are not in any scores, furthermore the change of Ca and osmolality could refer to survival. Although hypocalcaemia is common in Intensive Care unit, it is controversial if it is associated with the outcome [33]. Recent data confirm our findings with decreased prothrombin INR suggesting that it can be a predictive marker [12] and a previous investigation found association between osmolality and outcome in patients with acute coronary syndrome [13].

Evaluation of macrorheology remained in question. In whole blood viscosity a decreasing tendency could be detected in septic patients, while an increasing tendency in nonseptic patients, but no differences were found in plasma viscosity values. A previous study has described a similar result in WBV, but they found higher PV in sepsis [2]. Others have not found WBV different, but PV was lower in sepsis [21]. In spite of these differences in WBV, in our research it had no relationship to survival survival in all patients, but subgroup analysis showed a significant connection in nonseptic patients. Hyperviscosity reduces blood fluidity and causes microcirculatory failure [15], but in sepsis hyperkinetic circulation and slowly progressing anemia may prevent hyperviscosity syndrome [34]. It can explain that blood viscosity has a role only in nonseptic patients. Fibrinogen was elevated in sepsis and its increasing tendency was parallel with the elevation from day-1 to day-2 of red blood cell aggregation in nonseptic patients, but it affected the outcome neither in septic nor in nonseptic patients.

Red blood cell aggregation was increased in nonsurvivors, especially among nonseptic patients, and it correlated with ICU scores, suggesting that red blood cell aggregation can have a significant role in this patient group. Although no differences could be detected between septic and nonseptic patients, interestingly in nonseptic patients t½ shortened more over time than in septic patients, referring to a higher aggregation ability. In contrast, previous studies described that no relationship was found with outcome [2], but an association with sepsis and scores [32].

Several previous researches have explored impaired red blood cell deformability in sepsis and also revealed it as a marker of severity [24], it was lower than in nonseptic patients at higher shear stresses as our study indicated [9, 32]. Only one recent finding has reported a link between the worsening of deformability and mortality, even though they found it only in sepsis [9]. We observed a strong relationship between the deterioration of deformability and outcome in nonseptic patients, but not in sepsis. It can imply that in sepsis, where deformability is originally lower, further reduction does not have more serious consequences, but in nonseptic patients worsening can refer to the decreasing microcirculatory functions. Other explanation can be that in sepsis the profound microcirculatory alterations, increased permeability and capillary diameter, and the elevated vascular tone could hide the effect of deformability. Change of deformability might reflect the response to therapy or the capability to recovery. Patients, who could not maintain or increase the ability of red blood cells to deform, had higher risk to mortality.

Presence of sepsis has an explicit effect on survival and it is also a major component of ICU scores. Nevertheless, sepsis is largely different from nonseptic condition; there may be markers that refer to mortality only in septic or nonseptic patients. Hemorheological parameters had a dissimilar behaviour in septic and nonseptic patients and our results suggested that whole blood viscosity and red blood cell aggregation of nonseptic patients could be informative about ICU scores. The deterioration of red blood cell deformability may also add details to understanding mortality risk, although it is a dynamic parameter and data from the 2nd day are also needed, whereas scores use static variables from the 1st day. These parameters could be considered to be researched in a multicenter investigation.

The present study has some limitations. Firstly, it would provide more information if we performed measurements immediately after admission. Viscosity is affected by fluid intake; the different periods of time from admission until blood collection, the different amount of received infusion may limit our results. However, hemorheological parameters should be measured within 2 hours after blood sampling and it is not available any time during the day. Furthermore, it is questionable how to define the optimal time: earlier sampling will be inaccurate because of the therapy taken before the intensive care admission, later measurements will be affected by ICU therapies. Secondly, receiving transfusion can affect blood rheology. We decided to investigate a heterogeneous population including transfused patients. Our hypothesis was that survival depends on the properties of circulating blood, even if it is own blood or received. Thirdly, the relatively small sample size can limit the value of our observations and explanations. It can explain that hemorheological parameters had to be dichotomised, because survival analysis had poor results with continuous variables. A multicenter study is needed to confirm our findings and it could have clinical implications. Fourthly, these special parameters are not available in every laboratory and they are useful in one certain patient group, therefore it could only be optional in clinical practice.

In conclusion, our research suggested that Ca and osmolality can predict mortality in septic patients, and whole blood viscosity, red blood cell aggregation and the change of red blood cell deformability in nonseptic patients. Further investigations of microcirculatory alterations can help to understand pathophysiology of critical conditions and multicenter researches could evaluate the role of these parameters in estimating mortality risk.

Conflict of interest

The authors declare that there is no conflict of interest regarding the publication of this paper.

Footnotes

Acknowledgments

The present scientific contribution is dedicated to the 650th anniversary of the foundation of the University of Pecs, Hungary.

References

Alexy

, Wenby

R.B.

, Pais

, Goldstein

L.J.

, Hogenauer

, Meiselman

H.J.

An automated tube-type blood viscometer: Validation studies. Biorheology 42 (2005), 237–247.

Alt

, Amann-Vesti

B.R.

, Madl

, Funk

, Koppensteiner

Platelet aggregation and blood rheology in severe sepsis/septic shock: Relation to the Sepsis-related Organ Failure Assessment (SOFA) score. Clin Hemorheol Microcirc 30 (2004), 107–115.

Baskurt

O.K.

(2007), Mechanisms of blood rheology alterations. In: Baskurt

O.K.

, Hardeman

M.R.

, Ramling

M.W.

and Meiselman

H.J.

, (eds) Handbook of Hemorheology and Hemodynamics, IOS Press, Amsterdam, 2007, pp. 170–190.

Baskurt

O.K.

, Boynard

, Cokelet

G.C.

, Connes

, Cooke

B.M.

, Forconi

, Liao

, Hardeman

M.R.

, Jung

, Meiselman

H.J.

, Nash

, Nemeth

, Neu

, Sandhagen

, Shin

, George Thurston

and Wautier

J.L.

, New guidelines for hemorheologicallaboratory techniques. Clinical Hemorheology and Microcirculation 42 (2009), 75–97.

Baskurt

O.K.

, Gelmont

, Herbert

, Meiselman

H.J.

Red blood cell deformability in sepsis. Am J Respir Crit Care Med 157 (1998), 421–427.

Berliner

A.S.

, Shapira

, Rogowski

, Sadees

, Rotstein

, Fusman

, Avitzour

, Cohen

, Arber

, Zeltser

Combined leukocyte and erythrocyte aggregation in the peripheral venous blood during sepsis, An indication of commonly shared adhesive protein(s). Int J Clin Lab Res 30 (2000), 27–31.

Buchele

, De Backer

and Vincent

J.-L.

The microcirculation in critically-ill patients. Acta Clinica Belgica 62 (2006), 415–420.

Chierego

, Verdant

, De Backer

Microcirculatory alterations in critically ill patients. Minerva Anestesiologica 72 (2006), 199–205.

Donadello

, Piagnerelli

, Reggiori

, Gottin

, Scolletta

, Occhipinti

, Zouaoui Boudjeltia

and Vincent

J.-L.

, Reduced red blood cell deformability over time is associated with a poor outcome in septic patients. Microvascular Research 101 (2015), 8–14.

10.

Hardeman

M.R.

, Dobbe

J.G.

, Ince

The Laser-assisted Optical Rotational Cell Analyzer (LORCA) as red blood cell aggregometer. Clin Hemorheol Microcirc 25 (2001), 1–11.

11.

Hardeman

M.R.

, Goedhart

T.P.

, Schu

N.H.

Laser-assisted optical rotational cell analyser (LORCA), II: Red blood cell deformability: Elongation index versus cell transit time. Clin Hemorheol 14 (1994), 619–630.

12.

Helling

, Stephan

, Pindur

Coagulation and complement system in critically ill patients. Clin Hemorheol Microcirc 61 (2015), 185–193.

13.

Holtfreter

, Bandt

, Kuhn

S.O.

, Grunwald

, Lehmann

, Schütt

, Gründling

Serum osmolality and outcome in intensive care unit patients. Acta Anaesthesiol Scand 50 (2006), 970–977.

14.

Hurd

T.C.

, Dasmahapatra

K.S.

, Rush

B.F.

Jr., and Machiedo

G.W.

, Red blood cell deformability in human and experimental sepsis. Arch Surg 123 (1988), 217–220.

15.

Isbister

Hyperviscosity: Clinical Disorders. In Baskurt

O.K.

, Hardeman

M.R.

, Ramling

M.W.

and Meiselman

H.J.

, (eds) Handbook of Hemorheology and Hemodynamics, IOS Press, Amsterdam, 2007, pp. 371–392.

16.

Juneja

, Singh

, Nasa

, Dang

Comparison of newer scoring systems with the conventional scoring systems in general intensive care population. Minerva Anestesiol 78 (2012), 194–200.

17.

Jung

, Kelm

Evaluation of the microcirculation in critically ill patients. Clin Hemorheol Microcirc 61 (2015), 213–224.

18.

Kannel

W.B.

, D’Agostino

R.B.

, Belanger

A.J.

Fibrinogen, cigarette smoking, and risk of cardiovascular disease: Insights from the Framingham study. Am Heart J 113 (1987), 1006–1010.

19.

Keegan

M.T.

, Gajic

, Afessa

Severity of illness scoring systems in the intensive care unit. Crit Care Med 39 (2011), 163–169.

20.

Kesmarky

, Toth

, Habon

, Vajda

, Juricskay

Hemorheological parameters in coronary artery disease. Clin Hemorheol Microcirc 18 (1998), 245–251.

21.

Kirschenbaum

L.A.

, Aziz

, Astiz

M.E.

, Saha

D.C.

, Rackow

E.C.

Influence of rheologic changes and platelet–neutrophil interactions on cell filtration in sepsis. Am J Respir Crit Care Med 161 (2000), 1602–1607.

22.

Ledoux

, Canivet

J.-L.

, Preiser

J.-C.

, lle Lefrancq

and Damas

, SAPS 3 admission score: An external validation in a general intensive care population. Intensive Care Med 34 (2008), 1873–1877.

23.

Machiedo

G.W.

, Powell

R.J.

, Rush

B.F.

Jr. , Swislocki

N.I.

and Dikdan

, The incidence of decreased red blood cell deformability in sepsis and the association with oxygen free radical damage and multiple-system organ failure. Arch Surg 124 (1989), 1386–1389.

24.

Moutzouri

A.G.

, Skoutelis

A.T.

, Gogos

C.G.

, Missirlis

Y.M.

, Athanassiou

G.M.

Red blood cell deformability in patients with sepsis: A marker for prognosis and monitoring of severity. Clin Hemorheol Microcirc 36 (2007), 291–299.

25.

Muranaka

, Kunimoto

, Takita

, Sumino

, Nara

, Kuwano

, Murakami

Impaired blood rheology in critically ill patients in an intensive care unit. Journal of International Medical Research 34 (2006), 419–427.

26.

Nemeth

, Berhes

, Kiss

, Hajdu

, Deak

, Molnar

, Szabo

, Fulesdi

Early hemorheological changes in a porcine model of intravenously given E. coli induced fulminant sepsis. Clin Hemorheol Microcirc 61 (2015), 479–496.

27.

Ostergaard

, Granfeldt

, Secher

, Tietze

, Iversen

N.K.

, Jensen

M.S.

, Andersen

K.K.

, Nagenthiraja

, Gutiérrez-Lizardi

, Mouridsen

, Jespersen

S.N.

, Tonnesen

E.K.

Microcirculatory dysfunction and tissue oxygenation in critical illness. Acta Anaesthesiol Scand 10 (2015), 1246–1259.

28.

Piagnerelli

, Zouaoui Boudjeltia

, Vanhaeverbeek

and Vincent

J.-L.

, Red blood cell rheology in sepsis. Intensive Care Med 29 (2003), 1052–1061.

29.

Pöschl

J.M.B.

, Claude Leray

, Ruef

, Cazenave

J.P.

and Linderkamp

, Endotoxin binding to erythrocyte membrane and erythrocyte deformability in human sepsis and in vitro. Crit Care Med 31 (2003), 924–928.

30.

Powell

R.J.

, Machiedo

G.W.

, Rush

B.F.

Jr. and Dikdan

, Oxygen free radicals: Effect on red cell deformability in sepsis. Crit Care Med 19 (1991), 732–735.

31.

Powell

R.J.

, Machiedo

G.W.

, Rush

B.F.

Jr. , Decreased red blood cell deformability and impaired oxygen utilization during human sepsis. Am Surg 59 (1993), 65–68.

32.

Reggiori

, Occhipinti

, De Gasperi

, Vincent

J.-L.

and Piagnerelli

, Early alterations of red blood cell rheology in critically ill patients. Crit Care Med 37 (2009), 3041–3046.

33.

Steele

, Kolamunnage-Dona

, Downey

, Toh

C.H.

, Welters

Assessment and clinical course of hypocalcemia in critical illness. Crit Care 17 (2013), R106.

34.

Toth

, Kesmarky

Clinical significance of hemorheological alterations. In: Baskurt

O.K.

, Hardeman

M.R.

, Ramling

M.W.

, Meiselman

H.J.

, (eds) Handbook of Hemorheology and Hemodynamics, IOS Press, Amsterdam, 2007, pp. 392–404.

35.

Zimmerman

J.E.

, Kramer

A.A.

, McNair

D.S.

, Malila

F.M.

Acute Physiology and Chronic Health Evaluation (APACHE) IV: Hospital mortality assessment for today’s critically ill patients. Crit Care Med 34 (2006), 1297–1310.

36.

Zimmerman

J.E.

, Kramer

A.A.

, McNair

D.S.

, Malila

F.M.

, Shaffer

V.L.

Intensive care unit length of stay: Benchmarking based on Acute Physiology and Chronic Health Evaluation (APACHE) IV. Crit Care Med 34 (2006), 2517–2529.