Secondary Logo

Journal Logo

Perioperative medicine

Fluid replacement and respiratory function

comparison of whole blood with colloid and crystalloid

A randomised animal study

Fodor, Gergely H.; Babik, Barna; Czövek, Dorottya; Doras, Camille; Balogh, Ádám L.; Bayat, Sam; Habre, Walid; Peták, Ferenc

Author Information
European Journal of Anaesthesiology: January 2016 - Volume 33 - Issue 1 - p 34-41
doi: 10.1097/EJA.0000000000000251
  • Free



Blood loss during major surgery is associated with detrimental systemic and pulmonary effects. The best fluid replacement management following haemorrhage is one of the most polarising issues in anaesthesia and intensive therapy. Choosing the best fluid replacement from blood products, and various types of colloids and crystalloids, can be difficult. The safety of hydroxyethyl-starch (HES) has been brought into question in a recent meta-analysis,1 particularly in patients with increased capillary leakage. This and the risk of renal damage with HES,2,32,3 the appreciable cost of albumin and the defects of haemostasis induced by gelatin solutions have all reduced the available options.4 Crystalloids remain a rational option, but clinicians are reluctant to choose them because of the widespread belief that they move rapidly out of the circulation, a belief based on old studies that lacks an adequate evidence basis.5–75–75–7

We have recently shown that acute hypovolaemic shock and subsequent resuscitation with autologous blood affects respiratory mechanics.8 Although surgical procedures are associated with a more controlled but sustained blood loss that also requires fluid replacement therapy, the respiratory consequences of this have not been explored. The administration of blood products is often regarded as the gold standard therapy, with the main aim of maintaining oxygen transport capacity. However, no evidence-based data are available to allow a comparison of the changes in lung function between this consensual approach and goal-directed fluid therapy with colloids or crystalloids. This experimental study uses a novel animal model that mimics continuous, hidden surgical bleeding and replacement of the associated losses. We aimed to compare the effects of blood, colloid and crystalloid solutions on the flow resistance of the airways and on the viscoelastic properties of the respiratory tissues and then attempted to relate these changes to pulmonary oedema indices. We hypothesised that the respiratory effects of fluid resuscitation with blood would differ from those observed after colloid and crystalloid solutions.

Materials and methods

Ethical approval for this study (no. I-74-50/2012) was provided by the Experimental Ethics Committee of the University of Szeged, Szeged, Hungary, on 7 December 2012, and granted by the Animal Health and Welfare Office of the local authorities in Hungary (no. XIV/152/2013, Chairperson Cs. Farle) on 9 January 2013.

Animal preparation

Anaesthesia was induced with an intraperitoneal injection of 5% chloral hydrate (400 mg kg−1) in adult male Sprague Dawley rats. Tracheal intubation was achieved with a 16-gauge, polyethylene cannula (B. Braun Melsungen AG, Melsungen, Germany) after subcutaneous administration of local anaesthetic to ensure adequate analgesia around the surgical wound (lidocaine, 2 to 4 mg kg−1). The rats were then placed in a supine position on a heating pad and connected to a small animal ventilator (Model 683; Harvard Apparatus, South Natick, Massachusetts, USA), and mechanically ventilated with room air (70 breaths min−1, tidal volume 7 ml kg−1). A femoral vein was catheterised (Abbocath 22-gauge) for drug delivery and for fluid replacement. A femoral artery was cannulated (Abbocath 22-gauge) and attached to a pressure transducer (Model TSD104A; Biopac, Santa Barbara, California, USA) for continuous monitoring of mean arterial pressure (MAP), and to allow blood withdrawal, as part of the experimental protocol. The arterial pressure, ECG and heart rate (HR) were monitored continuously with a data collection and acquisition system (Biopac). Body temperature was kept in the 37 ± 0.5 °C range by using the heating pad.

Measurement of respiratory mechanics

The forced oscillation technique was applied in short (6-s long) end-expiratory pauses interposed in the mechanical ventilation sequence to measure the input impedances of the respiratory system (Zrs).9 Briefly, the ventilator was stopped at end-expiration and the tracheal cannula was switched from the ventilator to a loudspeaker-in-box system. The loudspeaker delivered a computer-generated small-amplitude (<1 cmH2O) pseudorandom signal (23 noninteger multiples between 0.5 and 21 Hz) through a 100-cm long, 2-mm internal diameter polyethylene tube. Two identical pressure transducers (model 33NA002D; ICSensors, Milpitas, California, USA) were used to measure the lateral pressures at the loudspeaker end (P1) and at the tracheal end (P2) of the wave-tube. The signals P1 and P2 were low-pass filtered (fifth-order Butterworth, 25-Hz corner frequency) and sampled with the analogue-digital board of a microcomputer at a rate of 256 Hz. Fast Fourier transformation with 4-s time windows and 95% overlapping was used to calculate the pressure transfer functions (P1/P2) from the 6-s recordings collected during apnoea. Zrs was calculated as the load impedance of the wave-tube.10 The input impedances of the tracheal tube and the connections were also determined and subtracted from each Zrs spectrum.

A model containing a frequency-independent resistance (R) and inertance (I), and tissue damping (G) and elastance (H) of a constant-phase tissue compartment11 was fitted to the Zrs spectra by minimising the weighted difference between the measured and the modelled impedance data. The tissue variables characterise the damping (resistive) and elastic properties of the respiratory system. Raw and Iaw represent primarily the resistance and inertance of the airways, as the contribution of the chest wall to these variables in rats is minor.12

Lung histology

After completion of the experimental protocol, the rats were euthanised with an overdose of intravenous (i.v.) pentobarbital sodium (300 mg kg−1). Midline thoracotomy was then performed and 4% formaldehyde was instilled into the right lung via the tracheal cannula at a hydrostatic pressure of 20 cmH2O after clamping of the left main bronchus near the bifurcation. The right lung was dissected and placed into 4% buffered formalin until further processing. After complete fixation, horizontal trans-hilar sections (perpendicular to the longitudinal axes of the lung from the hilum) were embedded in paraffin. Two 5 μm sections were prepared in each lung specimen and were stained with haematoxylin-eosin. Digitalised images were used to obtain the oedema index around randomly selected pulmonary vessels by dividing the lumen area by the total area of the pulmonary vessel (oedema cuff area + vessel lumen area). Histological images were analysed by the same investigator in a blind fashion and in a random sequence by using JMicroVision image analysis software (version 1.2.7).

Three to four tissue samples were dissected from the different lobes of the unfixed left lungs; these samples were weighed to establish the wet-to-dry weight ratio (W/D) as an index of the lung water content.

Experimental protocol

The rats were randomly assigned into one of the three protocol groups. The rats in Group B always received autologous heparinised blood (n = 8). Fluid replacement was performed with a colloid solution (HES 6% 130/0.4; Fresenius Kabi Deutschland GmbH, Bad Homburg v.d.H., Germany) in Group CO (n = 8), or with a crystalloid solution (NaCl 0.9%; B. Braun Melsungen AG, Melsungen, Germany) in Group CR (n = 9). The experiment began with standardisation of the lung volume history through the administration of a hyperinflation created by occluding the expiratory port of the ventilator once the animal had reached a steady-state condition (5 to 10 min after starting mechanical ventilation). The baseline respiratory mechanics were then established by measuring three or four reproducible Zrs data epochs. Next, haemorrhage was induced by the withdrawal of 5% of the estimated total blood volume13 via the femoral artery (Fig. 1). Three minutes later, another set of Zrs data was collected, including three individual measurements at 1-min intervals. The withdrawn blood was used for blood gas analyses (Cobas b221; Roche Diagnostics, Basel, Switzerland) to determine the haematocrit (Hct), pH and oxygen (paO2) and carbon dioxide (paCO2) partial pressures. The blood withdrawal and Zrs measurements were repeated once more in an identical manner. After completion of the first two steps of arterial haemorrhage, fluid replacement in accordance with the group allocation was performed by administering 5% of the total blood volume via the femoral vein. Three minutes after this manoeuvre, a set of Zrs data was recorded. The blood withdrawal/replacement procedure was repeated four more times, with the collection of Zrs data 3 min after each intervention. The total duration of resuscitation was around 90 min, with each step lasting approximately 7 min. Further arterial blood gas analyses were performed from the fourth and sixth blood samples. After completion of the measurement protocol, the lungs were processed for oedema assessment, as detailed above.

Fig. 1
Fig. 1:
Scheme of the experimental protocol. BG, assessment of arterial blood gas; BL, baseline; R1 to R5, fluid replacements; W1 to W6, blood withdrawals; Zrs, measurement of respiratory impedance data.

Data analysis

The scatters in the measured variables were expressed as SD values. The Kolmogorov–Smirnov test was used to test data for normality. Two-way repeated measures of analysis of variance (ANOVA), with the factors assessment time and group allocation, were used to assess the effects of blood loss and replacement on the respiratory mechanical and haemodynamic variables. The baseline respiratory mechanical assessments and oedema indices were compared by using one-way ANOVA. The Holm–Sidak multiple comparison procedure was applied to compare the different conditions (for repeated measures) or protocol groups (for independent groups). Correlation analyses between the variables were performed by using Pearson correlation tests. Statistical tests were carried out with the SigmaPlot software package (version 12.5; Systat Software, Inc., California, USA) with a significance level of P value of less than 0.05.


Rat body weights were not statistically different between the groups (344 ± 16.1 g for Group B, 320 ± 51.24 g for Group CO and 361 ± 20.7 g for Group CR). Table 1 summarises the baseline values of the respiratory mechanics for the three experimental groups. No statistically significant differences were detected in the variables reflecting the airway or tissue mechanics.

Table 1
Table 1:
Mean (SD) values of the airway resistance (Raw), tissue damping (G) and elastance (H) obtained under the baseline conditions in the three groups of rats

The arterial blood gas values obtained at the beginning, at the midpoint and at the end of the experimental protocol are presented in Table 2. In Group B, Hct did not change significantly throughout the protocol, whereas decreases in pH (P < 0.001) and paO2 (P = 0.011) were seen. Compared with autologous blood, fluid replacement with colloid solution resulted in a lower Hct (P < 0.001), while crystalloid administration led to significant reductions in both Hct (P = 0.010) and pH (P = 0.009). No significant difference in the changes in paO2 and pH was observed between the rats in Groups CR and CO. The decreases in Hct were more pronounced in Group CO than those in Group CR (P = 0.032).

Table 2
Table 2:
Mean (SD) values derived from arterial blood samples obtained at the first (W1), fourth (W4) and last (W6) withdrawal

Figure 2 depicts the changes in the airway and respiratory tissue mechanics relative to the baseline. Blood withdrawal resulted in a systematic lowering of Raw. Fluid replacement with colloid in Group CO restored the baseline value of Raw, whereas the Raw remained reduced following the i.v. administration of autologous blood in Group B (P = 0.005). The changes in Raw after the i.v. infusion of crystalloid solution in Group CR were less marked (P < 0.038), with less obvious elevations in Raw after the third fluid replacement manoeuvre. Repetitive increases in G were observed throughout the protocol (P < 0.001), with no statistically detectable differences between the protocol groups. H was elevated in all groups, with significantly greater changes in Groups CR (P = 0.005) and CO (P = 0.012) than in Group B.

Fig. 2
Fig. 2:
Changes in the airway (Raw: airway resistance) and tissue mechanics (G: damping, H: elastance) relative to the baseline (BL) during blood withdrawals (W1 to W6) and fluid replacements (R1 to R5) with autologous blood (Group B), colloid (Group CO) or crystalloid (Group CR). BV, total blood volume. * P < 0.05 vs. Group B within a condition, # P < 0.05 vs. Group CO within a condition.

The oedema indices obtained from the lung weights and from the histological analyses are seen in Fig. 3. The animals in both Groups CR and CO exhibited significantly greater wet-to-dry lung weight ratios (P < 0.001 for both); this was also manifested in the perivascular pulmonary oedema indices (P < 0.05 for both).

Fig. 3
Fig. 3:
Oedema indices obtained by relating the wet lung weight to the dry weight (left) and by relating the perivascular oedema area to the total vessel area on histological sections obtained in rats receiving autologous blood (Group B), colloid (Group CO) or crystalloid (Group CR). * P < 0.05.

The systemic haemodynamic changes for the three groups of rats are displayed in Fig. 4. The blood withdrawals caused MAP to decrease systematically, and it was restored to previous values by the fluid replacements, regardless of the group allocation. HR gradually increased in all groups of rats, with significant changes from R3, W3 and R2 in Groups B, CR and CO, respectively.

Fig. 4
Fig. 4:
Systemic haemodynamic variables during blood withdrawals (W1 to W6) and fluid replacements (R1 to R5) with autologous blood (Group B), colloid (Group CO) or crystalloid (Group CR). BV, total blood volume; HR, heart rate; MAP, mean arterial pressure; BL, baseline. * P < 0.05 vs. BL within a group.

The relationships between the wet-to-dry lung weight ratio and the relative change in H are presented in Fig. 5. Pooling of the data from the three protocol groups revealed significant correlations between the macroscopic oedema index and the increased stiffness of the respiratory system (r = 0.55, P < 0.01).

Fig. 5
Fig. 5:
Relationship between the changes in oedema index (wet weight/dry weight) and in respiratory elastance (H) in rats receiving autologous blood (Group B), colloid (Group CO) or crystalloid (Group CR).


Although recent studies have focused on the morbidity and mortality related to colloid or crystalloid use as fluid replacement therapy,1,2,141,2,141,2,14 the pulmonary effects of these solutions are mainly based on empirical investigations without firm evidence.5–75–75–7 Our experimental rat model was intended to mimic continuous insidious surgical bleeding and fluid replacement with solutions commonly used in clinical practice, and allowed us to assess mechanical properties specific for the airway and respiratory tissues following blood loss and replacement. The airway resistance decreased subsequent to the haemorrhage and remained low after fluid therapy with autologous blood, but returned to baseline following colloid and increased slightly after crystalloid. The respiratory tissues stiffened more markedly in the animals receiving colloid and crystalloid, with no difference in effect between these solutions. These adverse mechanical changes were also reflected in changes in the oedema indices determined by lung weight and by histology.

There has recently been an extensive debate concerning the best type and quantity of fluid replacement therapy following blood loss. Blood products are used to maintain the normal haemoglobin content of the circulation and ensure oxygen transport. Restoration of the circulatory blood volume by blood products has a beneficial effect on the preservation of the microcirculation with minimal morphological damage or ischaemic cell injury.15,1615,16 In common with previous findings, the blood loss in the present study led to bronchodilation, which is most probably due to the compensatory increase in thoracic gas volume and/or the elevated levels of circulatory catecholamines.8 Our findings add to what is known from a different model of haemorrhage that does not induce the severe hypovolaemia characteristic of hidden, leaky blood loss during major surgery.

Our results demonstrate that the Raw essentially remains lowered after administration of autologous blood. The lack of a complete recovery in airway tone could be attributed to the relaxation potential of heparin,17 but a comparison of heparinised and nonheparinised colloid solutions revealed no difference in their bronchial effects (data not shown), and the potential role of heparin can therefore be excluded. Alternatively, the depressed Raw may be attributed to the presence of bronchoactive mediators in the sequestered blood, with the particular importance of increased levels of adrenaline and noradrenaline in the withdrawn and subsequently re-administered blood.8 In contrast to autologous blood, colloid completely reversed the haemorrhage-induced bronchodilation. This illustrates the interaction between circulatory changes and airway mechanics, with recovery of the original airway geometry following restoration of normovolaemia. The increase in Raw following colloid administration may be attributed to a distension of the bronchial submucosal vessels and/or to oedema formation resulting in airway wall thickening, or an exudation into the airway lumen.18 A similar concept can be applied to the situation following crystalloid, the first administration of which fully reversed the decrease in Raw, when its entire volume was likely to remain in the vascular bed. This effect of the elevated intravascular volume may have been abolished in the rats of Group B due to the presence of catecholamines in the readministered autologous blood.

Following blood administration, there were slight, gradual increases in tissue viscoelasticity, which can be attributed to atelectasis and subsequent loss of lung volume induced by anaesthesia and mechanical ventilation in the supine position. This phenomenon was confirmed by the decrease in paO2, which indicates loss of alveolar surface available for gas exchange. An important finding of our study is the gradual, more marked impairment of respiratory tissue viscoelasticity in the animals receiving colloid or crystalloid solution (Fig. 2). This might be due to the different rheological properties of the fluids affecting the behaviour of respiratory tissue,19 or change in the colloid osmotic pressure related to haemodilution. As these adverse changes were also reflected in the oedema indices (Figs. 3 and 5), it is possible to anticipate that the accumulation of perivascular oedema will create stiffness in the compromised respiratory tissue. It is noteworthy that no difference was found between colloid and crystalloid treatments either in the changes in tissue mechanics or in the oedema indices. This suggests that, in terms of compromising lung tissue viscoelasticity and pulmonary oedema formation, these two solutions are equivalent, further supported by the lack of difference in the changes of blood oxygenation following the two fluid replacement regimens (Table 2). Studies that show little difference between colloid and crystalloid in effects on extravascular lung water, pulmonary leak index and lung injury score are in agreement.20–2220–2220–22 Our protocol covered only 90 min, and although there might be an agreement within this period, the prolonged effects would require further study.

An important methodological aspect of our protocol is related to the nature and the volume of the fluid administered. There are a variety of fluids used in resuscitation, but no consensus as to which is the best.23 Some meta-analyses suggest that albumin is well tolerated in the critically ill,24 although others do not recommend it because of lack of robust evidence that it reduces mortality.25 A recent international consensus promotes crystalloids over both HES and albumin solutions.23 Because we wished to compare the effect of three basic fluid replacement strategies, we deliberately selected HES as the colloid solution comparator. For the crystalloid solution, we chose isotonic normal saline rather than the hypotonic Ringer's lactate. There is an evolving debate on the pros and cons of balanced salt solutions over normal saline; the latter matches our other two fluid replacement strategies for osmolarity and was selected for a comparison with the slightly hypertonic HES 6% 130/0.4.26 With regard to the volume of crystalloid solution for fluid replacement, no evidence-based recommendations are available. The conventional view is that the volume of crystalloid to be administered should be three to four times the blood loss,27 but recent studies have questioned this, suggesting a ratio close to 1 : 1.2,7,282,7,282,7,28 As our interest was principally in the acute effects of fluid replacement, the volume administered for both solutions was the same as that chosen for the blood loss, 5% of the total blood volume. The similarity in MAP and HR between the protocol groups confirms this approach, and is in accord with the concept of goal-directed therapy.28,2928,29

One methodological limitation of our study is the use of total respiratory impedance data to assess pulmonary changes. Although Raw accurately reflects the flow resistance of the airways, the chest wall contributes significantly to the tissue variables G and H.12 Nevertheless, the viscoelastic properties of the chest wall exhibited negligible change following the induction of severe oedema with oleic acid.30 Therefore, we feel that our results probably reflect pulmonary changes; however, their magnitude may be somewhat underestimated due to the masking effect of the chest wall. Another methodological limitation is related to the species difference between small rodents and humans, necessitating caution in the extrapolation of our data to a clinical situation. Although rats have substantially higher Raw, G and H than humans, no major differences exist between mammalian species in the oscillatory mechanics apart from scaling factor.31

In summary, our results have provided experimental evidence of the dissociated changes in the airway and tissue mechanical properties following surgical-type bleeding and its treatment with autologous whole blood, colloid or crystalloid solution in a volume that fully restored MAP. Histological analysis and measurement of respiratory mechanics and gas exchange following blood loss and consecutive fluid replacement strategies revealed no differences between replacement with colloid and crystalloid. The two solutions demonstrated similar abilities to compromise the lung tissue viscoelasticity subsequent to mild perivascular oedema formation. These findings highlight the differences in behaviour of the respiratory system following fluid replacement with blood, colloid or crystalloid: a sustained bronchodilation is expected after the administration of autologous blood, without significant lung tissue changes, whereas colloids and crystalloids tend to restore the basal airway tone at the expense of deterioration in lung tissue viscoelasticity.

Acknowledgments relating to this article

Assistance with the study: the authors are indebted to Orsolya Ivánkovitsné Kiss for her invaluable assistance in the experiments. The authors thank József Kaszaki for his excellent advice, and Gabriella Varga for her help in the surgical preparation.

Financial support and sponsorship: this work was supported by a Hungarian Scientific Research Grant (OTKA K81179). It was supported by the European Union and the State of Hungary, cofinanced by the European Social Fund in the framework of TÁMOP 4.2.4.A/2-11-1-2012-0001 ‘National Excellence Program’ and 4.2.2.A-11/1/KONV-2012-0052.

Conflicts of interest: none.

Presentation: preliminary data were reported as an oral presentation at the annual congress of the European Respiratory Society in Munich, Germany, 6 to 10 September 2014.


1. Perel P, Roberts I, Ker K. Colloids versus crystalloids for fluid resuscitation in critically ill patients. Cochrane Database Syst Rev 2013; 2:CD000567.
2. Myburgh JA, Finfer S, Bellomo R, et al. Hydroxyethyl starch or saline for fluid resuscitation in intensive care. N Engl J Med 2012; 367:1901–1911.
3. Perner A, Haase N, Guttormsen AB, et al. Hydroxyethyl starch 130/0.42 versus Ringer's acetate in severe sepsis. N Engl J Med 2012; 367:124–134.
4. de Jonge E, Levi M. Effects of different plasma substitutes on blood coagulation: a comparative review. Crit Care Med 2001; 29:1261–1267.
5. McIlroy DR, Kharasch ED. Acute intravascular volume expansion with rapidly administered crystalloid or colloid in the setting of moderate hypovolemia. Anesth Analg 2003; 96:1572–1577.
6. Hahn RG, Drobin D, Stahle L. Volume kinetics of Ringer's solution in female volunteers. Br J Anaesth 1997; 78:144–148.
7. Gondos T, Marjanek Z, Ulakcsai Z, et al. Short-term effectiveness of different volume replacement therapies in postoperative hypovolaemic patients. Eur J Anaesthesiol 2010; 27:794–800.
8. Bayat S, Albu G, Layachi S, et al. Acute hemorrhagic shock decreases airway resistance in anesthetized rat. J Appl Physiol 2011; 111:458–464.
9. Petak F, Hantos Z, Adamicza A, et al. Methacholine-induced bronchoconstriction in rats: effects of intravenous vs. aerosol delivery. J Appl Physiol 1997; 82:1479–1487.
10. Franken H, Clement J, Cauberghs M, Van de Woestijne KP. Oscillating flow of a viscous compressible fluid through a rigid tube: a theoretical model. IEEE Trans Biomed Eng 1981; 28:416–420.
11. Hantos Z, Daroczy B, Suki B, et al. Input impedance and peripheral inhomogeneity of dog lungs. J Appl Physiol 1992; 72:168–178.
12. Petak F, Hall GL, Sly PD. Repeated measurements of airway and parenchymal mechanics in rats by using low-frequency oscillations. J Appl Physiol 1998; 84:1680–1686.
13. Lee HB, Blaufox MD. Blood volume in the rat. J Nucl Med 1985; 26:72–76.
14. Mutter TC, Ruth CA, Dart AB. Hydroxyethyl starch (HES) versus other fluid therapies: effects on kidney function. Cochrane Database Syst Rev 2013; 7:CD007594.
15. Onen A, Cigdem MK, Deveci E, et al. Effects of whole blood, crystalloid, and colloid resuscitation of hemorrhagic shock on renal damage in rats: an ultrastructural study. J Pediatr Surg 2003; 38:1642–1649.
16. Appel PL, Shoemaker WC. Evaluation of fluid therapy in adult respiratory failure. Crit Care Med 1981; 9:862–869.
17. Abraham WM, Abraham MK, Ahmed T. Protective effect of heparin on immunologically induced tracheal smooth muscle contraction in vitro. Int Arch Allergy Immunol 1996; 110:79–84.
18. Widdicombe J. Physiologic control. Anatomy and physiology of the airway circulation. Am Rev Respir Dis 1992; 146:S3–S7.
19. Mogensen ML, Steimle KS, Karbing DS, Andreassen S. A model of perfusion of the healthy human lung. Comput Methods Programs Biomed 2011; 101:156–165.
20. van der Heijden M, Verheij J, van Nieuw Amerongen GP, Groeneveld AB. Crystalloid or colloid fluid loading and pulmonary permeability, edema, and injury in septic and nonseptic critically ill patients with hypovolemia. Crit Care Med 2009; 37:1275–1281.
21. Pearl RG, Halperin BD, Mihm FG, Rosenthal MH. Pulmonary effects of crystalloid and colloid resuscitation from hemorrhagic shock in the presence of oleic acid-induced pulmonary capillary injury in the dog. Anesthesiology 1988; 68:12–20.
22. Hahn RG. Why are crystalloid and colloid fluid requirements similar during surgery and intensive care? Eur J Anaesthesiol 2013; 30:515–518.
23. Raghunathan K, Murray PT, Beattie WS, et al. Choice of fluid in acute illness: what should be given? An international consensus. Br J Anaesth 2014; 113:772–783.
24. Xu JY, Chen QH, Xie JF, et al. Comparison of the effects of albumin and crystalloid on mortality in adult patients with severe sepsis and septic shock: a meta-analysis of randomized clinical trials. Crit Care 2014; 18:702.
25. Patel A, Laffan MA, Waheed U, Brett SJ. Randomised trials of human albumin for adults with sepsis: systematic review and meta-analysis with trial sequential analysis of all-cause mortality. BMJ 2014; 349:g4561.
26. Ertmer C, Van Aken H. Fluid therapy in patients with brain injury: what does physiology tell us? Crit Care 2014; 18:119.
27. Puyana JC. Fink MP. Resuscitation of hypovolemic shock. Textbook of critical care. Philadelphia, PA: Elsevier Saunders; 2005. 1939–1940.
28. Hiltebrand LB, Kimberger O, Arnberger M, et al. Crystalloids versus colloids for goal-directed fluid therapy in major surgery. Crit Care 2009; 13:R40.
29. Doherty M, Buggy DJ. Intraoperative fluids: how much is too much? Br J Anaesth 2012; 109:69–79.
30. Barnas GM, Stamenovic D, Lutchen KR. Lung and chest wall impedances in the dog in normal range of breathing: effects of pulmonary edema. J Appl Physiol 1992; 73:1040–1046.
31. Bates JH, Irvin CG, Farre R, Hantos Z. Oscillation mechanics of the respiratory system. Compr Physiol 2011; 1:1233–1272.
© 2016 European Society of Anaesthesiology