PULMONARY FUNCTION IS BETTER PRESERVED IN PIGS WHEN ACUTE NORMOVOLEMIC HEMODILUTION IS ACHIEVED WITH HYDROXYETHYL STARCH VERSUS LACTATED RINGER'S SOLUTION

Abstract

Acute normovolemic hemodilution (ANH) has been proposed to decrease the need for allogenic blood transfusion. Consequently, great amounts of fluids are necessary to maintain hemodynamics during and after blood removal. The aim of this experiment was to evaluate the oxygenation, respiratory compliance, and lung structure during ANH performed with lactated Ringer's solution and hydroxyethyl starch (HES). Gas exchange, extravascular lung water, intrathoracic blood volume, serum osmolality, respiratory mechanics, and optical and electronic microscopy of lung biopsies were evaluated. Animals were randomized into three groups: CTL (control, n = 9), HES (HES 6% 200/0.5, n = 9), and LR (lactated Ringer's solution, n = 9). Animals in groups HES and LR underwent ANH to reach a preestablished hematocrit of around 15%. The removed blood was replaced with HES in a proportion of 1:1 and with lactated Ringer's solution 3:1. The LR group demonstrated a tendency for a marked time-dependence decrease in compliance (P = 0.013 in T2; P = 0.008 in T3) and in Pao₂/fraction of inspired oxygen (Fio₂) ratio (P = 0.033 in T2) as well as an increase in (A-a) Grad O₂ (P = 0.037 in T2). Extravascular lung water and intrathoracic blood volume did not present any significant variation among the groups. In contrast, serum osmolality presented a significant decline in animals hemodiluted with lactated Ringer's solution. Optical and electronic microscopy of lungs biopsies revealed moderate to serious collapses and basement membrane enlargement in LR group. In this kind of experimental model, ANH with 6% HES (200/0.5) seems to preserve lung structure better as evidenced by maintenance of oxygenation indexes and respiratory compliance when compared with that in the Ringer's solution hemodiluted group.

INTRODUCTION

Acute normovolemic hemodilution (ANH) has been used in an attempt to decrease the need for allogenic blood transfusion in the course of a variety of surgical procedures (¹). Nowadays, there is a widespread acceptance of lower hematocrits in both the operating room and the critical care unit (²). The volume resuscitation in critical situations, such as emergency surgery of traumatized and septic shock patients, may also provoke acute diminution in hematocrits (^{3, 4}). Many controversies exist concerning ANH such as the target hematocrit, the most appropriate fluid to be used (⁵), the risks incurred (⁶), its real benefits (^7-9), and how the anesthetic technique could influence the compensatory mechanisms (^{10, 11}). Its effects on the heart and brain have been extensively studied (^{10, 12}). However, the literature concerning its pulmonary effects is scarce (^13-16). Theoretically, ANH may affect lung function and gas exchange (¹⁵). The replacement of blood with large amounts of fluids during ANH may decrease plasma colloid osmotic pressure (COP) causing pulmonary edema (¹⁷).

The inverse relationship between plasma COP and transendothelial fluid flux is important to the maintenance of fluid balance in the lung, and the value of COP as an edema protective mechanism has been clarified (¹⁸). On the other hand, it has been demonstrated that ANH may improve the gas capability of the normal lung, as shown by increasing PaO₂ and lowering alveolar arterial oxygen gradient [(a-A)DO₂] in rabbits (¹⁵). These authors hypothesized that the improvement in arterial oxygenation during isovolemic anemia may be explained by regional lung VA/Q heterogeneity reduction. Conversely, in an experimental model and clinical investigation on acutely or chronically compromised lung, ANH resulted in a poorer gas exchange, lowering the arterial PaO₂ (^{14, 19}). More recently, in patients with chronic obstructive pulmonary disease undergoing one-lung ventilation who underwent mild ANH, Szegedi et al. (¹⁴) showed impairment in gas exchange. The same effect was not observed in patients with normal lung function. There is a suspicion that in chronic injured lungs, the NO mechanism involved in gas exchange amelioration during ANH may be compromised. We hypothesized that large doses of fluids used during hemodilution to replace blood could affect lung parenchyma structure and, consequently, gas exchange and respiratory mechanics. Because this matter has not been clarified in the literature, the purpose of this experimental model was to evaluate ANH effects on oxygenation, lung water, respiratory compliance, and lung structure by using two different fluids: lactated Ringer's solution and hydroxyethyl starch (HES; 200/0.5).

METHODS

Anesthesia and ventilation

After approval of the local Animal Investigations Committee, 27 Landrace × Large White crossbreed female pigs were studied. The animals (weighing 40.6 ± 7.2 kg) were restricted from food overnight but had free access to water. Pigs were premedicated with a combination of intramuscular ketamine (10 mg/kg), fentanyl (5 μg/kg), and midazolam (0.5 mg/kg). Anesthesia was induced with propofol (5 mg/kg); animals were intubated and connected to a mechanical ventilator (Galileo Hamilton Medical AG, Rhäzüns, Switzerland). A rigorous protocol consisting of volume-controlled ventilation, tidal volume set at 6 mL/kg, positive end-expiratory pressure (PEEP) at 5 cm H₂O, and an oxygen inspired fraction at 40% was kept in all animals. Respiratory rate was adjusted to maintain end-tidal CO₂ (Etco₂) between 35 and 45 mmHg. To avoid possible effects of the inhaled anesthetics on pulmonary mechanics, continuous intravenous anesthesia consisting of ketamine (5 mg/kg/h), fentanyl (10 μg · kg⁻¹ · h⁻¹), and pancuronium bromide (5 μg · kg⁻¹ · min⁻¹) as muscle relaxant was standardized for all animals. Bladder catheterization was performed for urinary volume control.

Dead space was estimated according to the formula Paco₂ − Etco₂/Paco₂ (²⁰). Oxygenation was evaluated by Pao₂/fraction of inspired oxygen (Fio₂) ratio, (a-A) DO₂, and intrapulmonary shunt (Qs/Qt) according to the standard formula (Qs/Qt = Cc′o₂ − Cao₂/Cc′o₂ − Cvo₂).

Monitoring and instrumentation

Hemodynamics-

Using electrocardiogram, heart rate was obtained directly from the monitor (IntelliVue MP40, Phillips, Boeblinger, Germany). Catheters were inserted through peripheral cut-downs in the left femoral artery and femoral and jugular right veins to carry out the hemodilution process and hemodynamics measurements. A pulmonary thermodilution and continuous Svo₂ artery catheter (model 744H-7.5F; Baxter Healthcare Corporation, Irvine, Calif) was positioned in the pulmonary artery trunk. Cardiac output and Svo₂ were continuously obtained through a cardiac monitor (Vigilance; Baxter Healthcare Corporation).

Extravascular lung water-

The extravascular lung water (EVLW) index was measured by transpulmonary thermodilution method. A 5-F thermistor-tipped catheter (Pulsiocath PV2015L20A; Pulsion Medical Systems, Munich, Germany) was placed in the femoral artery and connected to the IntelliVue monitor to obtain the EVLW index and intrathoracic blood volume (ITBV) by means of a specific software. Percutaneous continuous cardiac output was measured by the single indicator-transpulmonary thermodilution technique. Measurements were obtained by injections of 10 mL of cold saline solution, at a temperature inferior to 5°C, via the distal port of the central venous catheter placed in the internal jugular with subsequent detection by the thermistor embedded in the wall of the femoral artery catheter. The CO was calculated from the thermodilution curves according to the Stewart-Hamilton principle. The mean of three consecutive CO measurements was used. The percutaneous continuous cardiac output, using only one cold indicator, calculates the mean transit time and the exponential downslope time of the thermodilution curve. The product of CO times mean transit time is the intrathoracic thermal volume, whereas the product of CO times downslope time is the pulmonary thermal volume (PTV). Extravascular lung water is estimated as intrathoracic thermal volume minus the estimated ITBV (²¹).

Data provided correspond to the average of three measurements.

Respiratory mechanics

Respiratory mechanic parameters (i.e., inspiratory peak pressure, plateau pressure, total airway resistance, and static compliance) were obtained directly from the ventilator. Before the experiment, the ventilator transducers used to obtain respiratory mechanics were calibrated (Timeter RT 200; Allied Healthcare, St. Louis, Mo). Breathing circuit and tracheal tube compliance and resistance, respectively, were taken into account and subtracted from the data according to the conventional approach (²²).

Biochemistry and serum osmolality

Blood samples were drawn simultaneously from the pulmonary and arterial catheters for gas analysis. The arterial hematocrit was obtained with a microhematocrit centrifuge. Serum osmolality was measured at established points of the protocol by means of a osmometer (3300 Micro Osmometer; Advanced Instruments, Norwood, Mass).

Study design

Thirty minutes after anesthesia stabilization, basal measurements were obtained. At this point, the animals were randomized into three groups: control (n = 9), HES (n = 9), and lactated Ringer's (LR; n = 9). To calculate the amount of blood to be removed for hemodilution, the equation derived by Gross (²³) was applied: V_L = EBV × (Ht₀ − Ht_F)/Ht_ave, where V_L is the blood volume to be removed; EBV, exchangeable pigs' blood volume (80 mL/kg); Ht₀, baseline hematocrit; Ht_F, target hematocrit; Ht_ave, average hematocrit. Animals from groups HES and LR underwent ANH to reach a target hematocrit of around 15%. The hemodilution procedure was accomplished in an average of 30 min. Plasma expansion was simultaneously performed with the administration of HES 200/0.5 (Haes-Steril; Fresenius-Kabi, Campinas, Brazil) in a proportion of 1:1 or lactated Ringer's solution 3:1 (Table 1). After ANH, episodes of diminution in mean systemic arterial pressure greater than 20% of baseline were treated with additional fluid (lactated Ringer's solution or starch) according to the respective group. Core body temperature was monitored and maintained at 38°C with warmed IV fluids and warming lights. To minimize anesthesia effects in respiratory mechanics, a standard alveolar recruitment maneuver, as described, was performed before the collection of each respiratory mechanics and blood samples data. The ventilator knob was turned to continuous positive airway pressure mode, and a progressive PEEP was applied until a plateau pressure of 20 cm H₂O was reached. This plateau was maintained for 30 s, and the maneuver was repeated three times with a 1-min interval between them. Data were collected before blood withdrawal (T0) at the end of hemodilution (30 min, T1) and 1 h (T2) and 2 h after the end of hemodilution (T3). At this point, the chest cavity was open, and lung samples were taken from the posterior regions of the right medium and left diaphragmatic lobes from the three groups. Five samples were collected from each animal for optical and electron microscopy analyses. The lung samples prepared for optical analyses were kept in 10% formaldehyde solution, and 5-μm histological sections were stained with hematoxylin and eosin. The lung samples prepared for electron microscopy analyses were fixed in 2% glutaraldehyde in 0.1 mol/L phosphate buffer (pH 7.3) for 3 h at 4°C. The specimens were postfixed in 1% osmium tetroxide in 0.1 mol/L phosphate buffer (pH 7.3) for 2 h at 4°C. Dehydration was completed in ethanol, and the tissue was embedded in epoxy resin. One observer pathologist who was blinded to the groups analyzed the presence of alveolar edema, hemorrhage, and collapse in the prepared sample lamina using a Nikon optical microscope at a magnification of ×400. The observer qualified, in a subjective manner, the positive findings as discrete, moderate, or serious. The same observer analyzed the samples prepared for electron microscopy (×52,000). At least five sections were observed in each sample.

Table 1: Body weight and hydric balance in control and hemodiluted pigs

Statistical analyses

Weight, diuresis, removed blood, and infused volume were analyzed with the paired Student t test (P < 0.05). ANOVAs for repeated measures were applied to continuous variables to determine whether a significant overall difference occurred within or between the groups. If a difference was found, a Tukey test was used to assess its significance. Where appropriate, P < 0.05 was considered significant.

RESULTS

The three groups were comparable in weight. As expected, diuresis, blood removed, and infused volume were statistically different among groups (Table 1). Baseline physiological parameters were similar in the three groups (Table 2). No relevant differences were found in the hemoglobin levels between HES and LR, and these animals had been sufficiently hemodiluted to reach a target hematocrit of around 15%. These values remained stable throughout the procedure. Acute normovolemic hemodilution resulted in a significant decrease in sodium in group LR (P = 0.007 in T3). As expected, serum osmolality in the LR group revealed a significant decrease after hemodilution (Table 2). As predicted, ANH resulted in an increase in cardiac index in both groups; however, in the HES group, it was higher (P < 0.001 in T1). A significant increase occurred in MAP pressure in the LR group at 1 and 2 h after the end of ANH (P = 0.028; P = 0.005). The initial values for inspiratory peak pressure and plateau pressure were considered statistically similar for the three groups. Group LR exhibited a tendency toward diminished lung compliance when compared with the control group (P = 0.013 in T2; P = 0.008 in T3) (Fig. 1). The LR group exhibited impairment in the PaO₂/FIO₂ (P = 0.033 in T2) index, thus recovering toward the initial values at the end of the experiment (Fig. 2A). Differently from animals hemodiluted with HES, ANH resulted in a certain trend toward an increased Qs/Qt and to a significant increase in dead space in the LR group (P = 0.041 in T2; P = 0.033 in T3) (Fig. 2, B and C).

Table 2: Systemic, pulmonary hemodynamic, and pulmonary mechanics

Fig. 1: Mean and standard error of pulmonary compliance in CTL (control animals), HES (hemodiluted with HES), and LR (hemodiluted with lactated Ringer's solution). *A significant compliance decrease occurred in the LR group in time 2 (P = 0.013) and time 3 (P = 0.008) when compared with that in the control group.

Fig. 2: Mean and standard error of Pao₂/Fio₂ ratio, Qs/Qt, and dead space in CTL, HES, and LR. ^#A significant decrease in Pao₂/Fio₂ in the LR group in time 2 (P = 0.033). *A significant dead space increase occurred in the LR group in time 2 (P = 0.041) and time 3 (P = 0.033) when compared with that in the control group.

Extravascular lung water and ITBV remained relatively unchanged in the three groups during the protocol established points to data collection. As pointed out in Figure 3, the basal values of EVLW were 10.2 ± 0.70, 9.5 ± 0.50, and 9.2 ± 0.71 in the control, HES, and LR groups, respectively, and at the end of procedure, the values were 9.7 ± 0.63, 9.7 ± 0.70, and 8.3 ± 0.60.

Fig. 3: Mean and standard error of EVLW and ITBV in CTL, HES, and LR.

Histopathology analysis

Regarding histopathology analysis (Figs. 4A and 5, A and B), no abnormalities were found in the control and HES lung biopsies. Interestingly, optical microscopy revealed moderate to serious collapses of lung parenchyma in the LR group (Fig. 4B), and electron microscopy showed basement membrane enlargement in the same group (Fig. 5C). With the exception of atelectasis in LR group, no evidences of abnormal lung fluid accumulation were observed in the optical microscopy in all specimens analyzed.

Fig. 4: Representative photomicrographs obtained from lungs of HES (A) and LR where an intense collapse is evidenced (B). Hematoxylin and eosin stain, original magnification ×400. Alveolar collapse (%) in CTL, HES, and LR (C).

Fig. 5: Electron microscopy obtained from lungs of CTL (A), HES (B), and LR (C); 1 = alveolar lumen, 2 = type I epithelial cell, 3 = basement membrane, 4 = capillary endothelial cell, 5 = capillary lumen, 6 = pinocytotic vesicle, 7 = erythrocyte. In C (group LR), an enlargement of basement membrane is evidenced. Original magnification ×52,000.

DISCUSSION

To our knowledge, this is the first study comparing the pulmonary effects of ANH with two different solutions: LR and HES 200/0.5. The principal results of this investigation show that animals hemodiluted with HES had respiratory compliance, airway resistance, and oxygenation indexes expressed by the Pao₂/Fio₂ ratio less affected when compared with those hemodiluted with lactated Ringer's solution. More impressively, lung structure specimens analyzed by optical and electron microscopy revealed extensive areas of alveoli collapse and basal membrane enlargement in the LR group. Surprisingly, even in the animals that presented significant modifications in respiratory compliance and gas exchange, the EVLW remained unaltered when compared with the control group. Transpulmonary thermodilution used in our study has been proposed as an alternative technique to estimate EVLW and hence to quantify pulmonary edema (²¹). This method has been shown to compare favorably with the double indicator (thermo-dye) dilution technique (²⁴) and with the ex vivo gravimetric method (²⁵).

Lung edema is defined as the abnormal accumulation of fluid in the extravascular space of the lung. However, hypoxemia and reduced lung compliance are not specific of fluid excess and, hence, correlate only very weakly with the amount of EVLW (²⁶). In this way, a quantitative assessment of extrapulmonary water was used in this study using dilution techniques. According to Michard and colleagues (²¹), the amount of pulmonary edema, the levels of tidal volume and PEEP, and the severity of lung disease may affect the accuracy of EVLW estimation. According to them, edematous lung areas may compress pulmonary vessels and enhance pulmonary vasoconstriction, both factors that may reduce pulmonary blood volume, and hence lead to overestimation of ITBV and underestimation of EVLW (²¹). In our study, the unchanged EVLW during hemodilution is coincident with the microscopy findings that did not reveal a grossly pulmonary edema in the analyzed lung specimens, just an extensive alveolar collapse in LR group. The extensive alveoli collapse accompanied by significant variations in respiratory compliance, observed in LR group when compared with starch-treated group, is difficult to explain. To prevent the initiation of ventilator-induced lung injury, transpulmonary pressure must be kept within the physiological range. In our protocol, we established rigid rules to avoid the impact of mechanical ventilation on lung structures. However, microvascular stress seems to be a potent cofactor in the development of pulmonary edema and lung damage resulting from an injurious pattern of ventilation (²⁷). An increase in cardiac output, as it occurs during ANH process, may be associated with an increased prealveolar microvascular pressure and a higher vascular pressure gradient across the lung. This abnormal pattern of flow may contribute to additional lung lesion when high inflation pressures are in use. Our animals, hemodiluted with lactated Ringer's solution, presented higher pulmonary artery pressures when compared with HES group. It remains speculative if lung alterations observed in LR group were caused by an increased prealveolar microvascular pressure, provoked by low osmolarity, or both.

The decrease in respiratory compliance in the LR group may be explained by alveolar atelectasis. The lungs were kept inflated intermittently by the same tidal volume during all the time because of the use of volume-controlled ventilation modality. As shown in Table 2, the plateau pressure in LR group increased progressively after the hemodilution (T2 and T3), reflecting extensive collapse of alveolar space as corroborated by optical microscopy (Fig. 3B).

Unfortunately, the literature concerning ANH pulmonary effects is scarce. Cooper and colleagues (²⁸) showed an appreciable increase in parenchymal lung water content evidenced by gravimetric and histological evaluations in dogs hemodiluted with lactated Ringer's solution. They were not able to detect any increase in EVLW by double indicator hemodilution method. Histologically, the water was exclusively present in peribronchial spaces. These effects were partially reversed by restoration of oncotic pressure by albumin (²⁸). We hypothesized that the large amount of fluids, three times more Ringer's solution than starch necessary during blood replacement, may have contributed to the extensive alveoli collapse. Its extensive collapse was probably caused by an initial stage of lung edema undetectable by EVLW. As we know, pulmonary edema occurs when safety mechanisms of the lung are besieged by either high transvascular pressure gradients, as in cardiogenic edema, or increases in the microvascular permeability to solutes, as in acute lung injuries. In their experimental model of pulmonary edema provoked by high pressure or by altering permeability, Staub and colleagues (²⁹) described the following hypothesis to explain the interrelation of atelectasis and edema. If the capillaries are not leaky and surface tension is relatively high, then the alveoli collapse to the atelectatic state. If the capillaries are very leaky and surface tension is relatively low, then it will fill with fluid at normal volume (²⁹). Our LR group was similar to the first situation described-increase in hydrostatic pressure accompanied by low oncotic pressure occasioned by large amount of fluids. During ANH process, as probably also occurs during large fluid infusion with nonproteic fluids during surgery or hypovolemic resuscitation, an initial unrecognized pulmonary edema by available methods at bedside may occur.

As happened in this study, a significant diminution in serum osmolality in LR hemodiluted animals may have had a marked contribution to pulmonary findings. The importance of colloid-osmotic pressure, as an edema protective mechanism, should be also considered. A 50% reduction in COP increases lymphatic flow by fourfold (¹⁷); lung water increases more substantially in hypoproteinemic dogs (³⁰), particularly during crystalloid administration (^31-34).

Regarding gas exchange, most of the studies in this area focus on the comparison of the ANH influence on normal and previous impaired respiratory function. Isovolemic anemia results in improved gas exchange in animals with normal lungs, but in relatively poorer gas exchange in those with whole-lung atelectasis. In normal lungs, Deem and colleagues (¹⁵) showed an increase in arterial Po₂ in hemodiluted rabbits. Conversely, the same authors (¹⁹) showed that arterial oxygenation in rabbits with an atelectatic lung provoked by selective bronchial intubation was impaired during ANH. Intrapulmonary shunt (Qs/Qt) was measured by using blood gas analysis and by quantification of the percentage of blood flow to the collapsed left lung using fluorescent microspheres. They hypothesized that an increase in NO during ANH has attenuated the hypoxic pulmonary vasoconstriction causing disarrangement of ventilation to perfusion matching, thus worsening the oxygenation. Deem et al. (¹⁶) reported that ANH improved oxygen exchange in rabbits with lung injury induced by gas embolism, but they did not find a clear mechanism for this improvement. A recent clinical investigation (¹⁴) reported that even mild ANH significantly decreased arterial oxygenation during one-lung ventilation in patients with chronic obstructive pulmonary disease. However, contrary to all these results, because our pigs had prior normal lung function, we expected, according to the literature, that ANH could promote an increase in arterial Pao₂. However, in the LR group, a significant decline in Pao₂ was observed at the end of ANH accompanied by an increase in shunt and dead space. We assumed that the extensive alveoli collapse may have acted as an acute injury, and ventilation perfusion matching is worsened by ANH as supported by previous studies (^{14, 19}). Considering the HES group, an overall improvement in gas exchange was expected once no lung injury was evidenced by microscopy. Nevertheless, gas exchange and respiratory mechanics did not vary during hemodilution, and the values were comparable to those in the control group. We could not find a plausible explanation for these data. As anticipated, after ANH, both groups had an elevation in cardiac index, significantly higher in the HES group (Table 2). The increment in intrapulmonary shunt may be explained by this increase in blood flow through the lungs. However, it is interesting that the intrapulmonary shunt did not accompany this increase in pulmonary flow as observed in the HES group. This finding corroborates the fact that altered pulmonary structure, as seen in the LR group, was more important than hemodynamic mechanisms to explain the increase in intrapulmonary shunt in this group. The basement enlargement revealed by electron microscopy is probably due to a decrease in osmolality (³⁵) associated with an increased capillary hydrostatic pressure (Fig. 4C). The alveoli collapse can explain the decreased compliance and oxygenation impairment in LR group. Possibly, according to the literature, the anemia leading to increased NO availability has attenuated the protective hypoxic pulmonary vasoconstriction, thus contributing to ventilation/perfusion mismatch in LR group.

A weakness in the method we used for ANH may be explained by the difference in hemoglobin between groups. This difference may reflect fluid shifts between compartments. Indicator dye dilution or dilution of radiolabeled red blood cells should be more precise to assure the amount of fluid remaining in the intravascular space after blood exchange during ANH. However, these methods are unsuitable for studies of non-steady-state conditions because they require an unchanged blood volume for a period of 30 to 40 min to be accurate (³⁶). Nevertheless, to maintain arterial pressure after ANH, a continuous infusion of fluids was necessary, mainly in LR group. Hemoglobin dilution method may also indicate changes in blood volume, but it has been questioned considering its precision being also unsuitable for our protocol (³⁷). Otherwise, 1 mL of blood loss replaced with 3 mL of balanced salt solution may be explained by studies of fluid kinetics (³⁸). Because of extravascular redistribution of crystalloid solutions, the volume infused must exceed the volume of blood withdrawn by a factor of at least 3. Colloid has the advantage of intravascular retention, which results in smaller volume replacement requirements. Thus, the amount administered was approximately equal to the volume of blood withdrawn. It is beyond our study to affirm that all kinds of colloids, including albumin, dextran, and gelatins, could present the same effect as we observed with HES. In fact, colloids are not equal in safety and pharmacokinetics properties. In their extensive review, Barron and colleagues (³⁹) suggested the need to consider the contrasting safety profile of colloids in clinical decision making.

CONCLUSION

In conclusion, hemodilution based in large crystalloid infusion resulted in the impairment of oxygen exchange and worsening of respiratory mechanics in the LR group. Supported by results achieved from this experiment, it can be assumed that ANH with HES is capable of maintaining pulmonary mechanics and oxygenation performance better than that with lactated Ringer's solution. It can be partially explained through the documented collapse and basal membrane alveoli edema that occurred in this group. Our message, considering the limitations of this being an experimental study, is to inform the apparent pulmonary risk related to great infusions of crystalloids solutions to replace blood.