Pulmonary Blood Flow Increases in Damaged Regions Directly after Acid Aspiration in Rats

ACID aspiration is a major cause of acute lung injury (ALI) and acute respiratory distress syndrome.^1,2  Approximately 15–16.5% of patients with acid aspiration may develop ALI,^3,4  and the incidence increases with further predisposing conditions (e.g., sepsis or trauma) and risk factors (e.g., chemotherapy or obesity).⁵  The type of aspirate can vary from fluid content (liquid, blood, or secretions) to food particles. Aspiration can cause an initial episode of pneumonitis, which may be complicated by subsequent bacterial pneumonia.⁵  ALI pneumonitis secondary to aspiration carries a mortality rate of 44%.⁶ 

The induction of pulmonary leukocyte sequestration and the disruption of capillary–alveolar barrier function were shown to be responsible for the development of pulmonary edema in animal models of acid-induced ALI.^7–10  In addition, polymorphonuclear leukocyte migration is initiated via diapedesis through the alveolar capillaries in response to inflammation.¹¹  Although aspiration injury has been investigated extensively, its impact on spatial pulmonary blood flow (PBF) during the early phase of injury has not been carefully evaluated. Data regarding very early pathophysiological changes are highly desirable, as the clinician must react within minutes after an aspiration event.

Experimental interventions that have been suggested to treat acid-induced lung damage include intravenous hyperosmolar solutions,⁷  hypertonic saline solutions,¹²  and pentoxifylline.¹³  Studying the regional blood flow pattern after an aspiration event could contribute to the development of novel therapeutic approaches, such as drug targeting. In the current study, we characterize the spatial PBF shortly after acid aspiration using radiolabeled microspheres and positron emission tomography (PET). A preliminary analysis of portions of this study has been previously reported in abstract form.¹⁴ 

After receiving approval from the local Animal Care Committee (Landesdirektion Dresden, Saxony, Germany), 18 male Wistar-Unilever rats (293–500 g) were included in this study. We anesthetized the rats using a combination of desflurane (7–12%), a single intraperitoneal injection of ketamine (75 mg/kg), and 1% lignocain infiltrated into the operative field. Catheters were introduced into the femoral artery (0.9-mm Leader catheter; Vygon, Ecouen, France) and femoral vein (0.8-mm Umbilical Vessel Catheter; Tyco Healthcare, Tullamore, Ireland) to measure arterial blood pressure and heart rate (Component Monitoring System; Hewlett-Packard Model 54S; Saronno, Italy), to obtain arterial blood samples for the blood gas analyses (ABL 50; Radiometer, Copenhagen, Denmark), to measure hematocrit, and to provide intravenous infusions. The spontaneously breathing animals were allowed to stabilize for 30 min after preparation and were then randomly assigned to either the control (n = 6) or the injury group (n = 12).

ALI was induced via the instillation of 0.1 m HCl (0.4 ml/kg body weight) through a tracheostomy tube with a catheter inserted to maximum depth, causing selective injury to either the left or right lung at random. The control group received no treatment through the catheter.

We measured physiologic parameters (blood gases and hemodynamics) at baseline and 10 min after the infusion of the radiolabeled microspheres. Human serum albumin microspheres with a mean diameter of 20 µm were labeled with Gallium-68 (⁶⁸Ga), as previously described,¹⁵  which yielded injectable ⁶⁸Ga human serum albumin microspheres (Ga-68 microspheres).

Spatial PBF was determined using a small animal PET scanner (microPET^® P4; Siemens Preclinical Solutions, Knoxville, TN). After a 10-min transmission scan, a 1-min bolus infusion of Ga-68 microspheres was initiated, and then a 30-min emission scan was performed. In 6 of the 12 injured animals, we performed a second transmission scan immediately after the 30-min emission scan to measure the potential influence of local density changes on the PET volume data and calculate altered PBF after injury. PBF differences between the left and right lobe were compared using different attenuation corrections (determined by transmission before and after the emission scan). The emission scans were corrected for decay and attenuation, and the frames were reconstructed using Ordered Subset Expectation Maximization that was applied to 3D sinograms with 14 subsets, six Ordered Subset Expectation Maximization 3D iterations, two maximum a posteriori iterations, and a 0.05 β-value for smoothing.

The degree of injury was assessed through lung density measurements, which were obtained from 10-min computer tomography (CT) scans of the entire lung acquired at baseline and at the end of the experiment using a micro-CT scanner (SkyScan 1178 micro-CT system; SkyScan, Kontich, Belgium) with the following specifications: 360 degrees of scanning, averaging two frames; 83-µm pixels; 85-mm field of view, scanning time of 9 min and 7 s; and a voltage of 65 kV and 615 µA. After the final measurement, the animals were sacrificed, and the lungs were extracted for postmortem analysis (fig. 1).

The lungs were removed at continuous airway pressures equal to an intrabronchial pressure of 7 cm H₂O and preserved for microsphere and tissue analyses. After extraction, the lungs were immediately frozen at −50°C using isopentane/dry ice and stored at −64°C. Six specimens were removed, embedded in tissue-freezing medium (Jung; Leica Microsystems, Wetzlar, Germany), and cut into 20-µm slices of the entire transverse plane (apical: 0.5 cm below the upper apex; middle: at the tracheal bifurcation level; basal: 1 cm above the most basal lung tissue) using a microtome (Leica CM 3000 cryostat, Leica Microsystems) at −20°C. Microscopic photographs were taken (AxioImager.A1) and processed (Axiovision, AxioVs40 V4.8.1.0; Carl Zeiss Imaging Solutions, Jena, Germany). After the addition of the mosaic pictures, complete images of the selected lung layer were obtained. The entire representative transverse planes of the apical, middle, and basal lung tissues were analyzed to determine the concentration of microspheres present. Diffuse alveolar damage (DAD) was evaluated by an expert blinded to the protocol using a weighted histological score.¹⁶  Values from 0 to 3 (0 = normal; 1 = slight effect; 2 = intermediate effect; and 3 = severe effect) represent the severity of each feature (alveolar edema, interstitial edema, microhemorrhage, inflammatory infiltration of polymorphonuclear neutrophils and macrophages, microatelectasis, and alveolar overdistension). The extent of damage in each sector was recorded as follows: 0 = no damage; 1 = up to 25% damage; 2 = 25–50% damage; 3 = 50–75% damage; 4 = 75% to approximately complete damage; and 5 = complete damage. When multiplied, the severity of the feature and the extent of damage produced values between 0 and 15, which were averaged for each region. Finally, we performed scanning electron microscopy (LEO S430; Carl Zeiss, Oberkochen, Germany) on representative lung specimens.

Computer tomography scans were coregistered to the PET scans, and regions of interest (ROI) were defined (fig. 2). PBF was calculated by analyzing the pulmonary distribution of Ga-68 microspheres within the ROIs (right and left lungs, ROI_specific).¹⁵  ROI_specific was defined by placing a spherical ROI of 0.24 cm³ over the area with the greatest Ga-68 microsphere concentration. An identical ROI_specific of the same volume was then placed on the contralateral lung in an anatomically corresponding position (fig. 2) and layered over the coregistered second CT scan of the animal to assess lung density and to further calculate the correlation between PBF and lung density within this ROI. PBF for the entire lung and for ROI_specific was analyzed as a ratio of right to left side for the control animals and for the injury to uninjured side for the aspiration group.

The standardized uptake value is defined as the Gallium-68 microspheres uptake in the selected lung regions (apical, medial, and basal ROIs for both groups) scaled by the administered activity and by the body weight of the rats studied, and the values were decay corrected with respect to the Gallium-68 half-life. To minimize recovery effects, we used maximal Gallium-68 microspheres uptake values to calculate maximal standardized uptake value.

All data are presented as the medians values 25th and 75th percentiles (median [25%; 75%]), unless otherwise specified. A nonparametric Wilcoxon two-tailed, matched-pairs signed-rank test was used for data comparison within groups, and an two-tailed, unpaired t test with Welch correction was used when appropriate (GraphPad Prism 5.02 for Windows; GraphPad Software Inc., San Diego, CA). Statistical significance was assigned for P values less than 0.05.

Animals from the control and aspiration groups (weighing 340 g [313; 381] and 354 g [333; 460], respectively) did not differ in terms of the physiological values (table 1) recorded at baseline. Acid aspiration generated major changes in the injury group, including deterioration of blood gases and significantly reduced arterial pressure (table 1). Although arterial oxygen tension, arterial pH, and arterial oxygen saturation declined in the injury group, the arterial carbon dioxide tension, heart rate, and respiratory rate remained unaffected. The physiological variables for animals in the control group were stable, with no evidence of blood loss or changes in body temperature during the experiment. Separate analysis of gas exchange and respiratory rate, heart rate, and mean arterial pressure revealed no differences within the injury group between the left- and right-side injured animals.

The PBF data are represented for different regions: right lung, left lung, ROI_specific of the injured and uninjured lung and for three transversal planes (apical, middle, and basal) of each lung.

PBF ratio of the right to the left lung for the control group was not significant different from the PBF ratio of the injured to the uninjured lung for the injury group (fig 3; P = 0.16). The PBF ratio of ROI_specific of the injured to uninjured side was 1.2 (1.1;1.3). This PBF ratio for ROI_specific was significantly higher in the injury group than the PBF ratio for ROI_specific of right to left lung of the control group (fig. 3; P < 0.0001). This was a result of a higher PBF on the injured side (0.33 [0.29; 0.34]) compared with the uninjured side (0.26 [0.29; 0.34]; P = 0.001).

The distribution of the regional PBF was not different within the control group, when comparing PBF data from the left and the right lungs. The group that had acid instilled into the right lung had no statistically significant increase in PBF to the entire right lung. However, there was a statistically significant increase in PBF to the left-acid instilled lungs. The PBF for ROI_specific was significantly higher in the injured lungs (table 2).

The PBF ratio of right to left lung for the control group was significantly different from the PBF ratio of injured to uninjured lung in the injury group for all chosen ROIs (fig. 4). Moreover, the difference between the ratios appeared to progress from the apical to the basal regions, indicating a slightly increased blood flow along the cranial to caudal axis in injured regions. Animals that underwent acid aspiration showed higher ipsilateral maximal standardized uptake values for the injured side compared with the contralateral side (for all regions in the right injury group and the middle and basal ROIs in the group with left injuries). Left lung-injured animals demonstrated significant differences in PBF on both lungs compared with the control group, whereas an injury on the right lung caused differences in PBF only in the injured lung for each ROI compared with the control group (fig. 5).

PBF emission scans reconstructed with transmission scans obtained either before or after emission scans (see figure, Supplemental Digital Content 1, https://links.lww.com/ALN/A961) confirmed that temporal changes in lung density did not influence the detection of blood flow effects. Absolute differences in percent increase of PBF between the injured and the uninjured lung resulted in 1.2–3.3 (2.0 [1.4; 2.6]) (see table, Supplemental Digital Content 2, https://links.lww.com/ALN/A962).

Reconstruction of the PET image by using the transmission scan before the emission resulted in a PBF difference (injured and uninjured lung) of 19.3% [10.5; 28.8], whereas the calculation based on the transmission scan after emission resulted in a 19.7% [12.4; 29.3] difference. The absolute difference (2%) is much smaller than the increase in PBF difference (19.3%, 19.7%; P < 0.001), thus representing 10% of the measured PBF increase of the injured lobe.

All rats displayed a relatively larger imaged right lung volume (2.6 cm³ [2.3; 3.8]) compared with left lung volume (2.0 cm³ [1.7; 2.4]; P = 0.02). After aspiration, changes in density varied interindividually in size and magnitude and were observed exclusively in the middle and basal regions. Compared with the baseline CT scan, there were areas of increased density in the CT images after aspiration in both the left lung (n = 6; size: 1.00 cm³ [0.65; 1.41]) and the right lung (n = 6; 1.18 cm³ [0.94; 1.3]) that represented 59% [43; 79] of the left lung, 48% [36; 61] of the right lung, and 23% [17; 29] of the entire imaged lung.

Lung density on the injured side within the ROI_specific was found to range from −100 to 12.95 Hounsfield Units in nonaerated areas.

The fraction of mean density ratio (injury/reference side of ROI_specific) was 1.3 [1.2; 1.5].

This demonstrates that the injury increased the density of studied lung region. The same region had also a higher PBF injured/uninjured ratio. However, the comparison of the PBF and density ratios showed that they did not correlate (Pearson r: 0.28; P = 0.37).

The DAD score values were slightly increased for the middle and basal ROIs (see figure, Supplemental Digital Content 3, https://links.lww.com/ALN/A963). The cumulative DAD scores were different in the basal ROI (fig. 6).

We found microspheres throughout the pulmonary vascular bed, even in acid-affected regions (fig. 7). Comparing microsphere numbers in the injured and the uninjured sides of the lung, we counted 15 [9; 17] and 14 [11; 23] microspheres for the apical layer, 27 [15; 35] and 16 [9; 27] microspheres for the middle layer, and 21 [17; 29] and 11 [10; 20] microspheres for the basal layer. These findings did not reach statistical significance.

Using the standardized acid aspiration model, we have shown effects of ALI on PBF in anesthetized, spontaneously breathing rats using small animal imaging techniques.

Our findings indicate that injured regions demonstrate increased PBF 10 min after acid aspiration and increased density in association with impairments in physiological variables within 50 min after injury.

Although the acid aspiration technique has been standardized, the distally administered HCl may have been distributed unevenly. The amount of HCl delivered was limited to avoid dissemination to the contralateral lung. It remains unclear how well this model imitates aspiration injury in humans due to the absence of particulates, bacterial products, and cytokines.¹⁷ 

Critique of the PET technique used here has been previously discussed.¹⁵  The data acquisition time for the PET scan was adjusted to obtain an optimal activity count rate. Changes in lung density during observation did not significantly influence the measured PBF derived from the volume data, which were reconstructed with attenuation correction using transmission scans before the emission scan. A caveat was our inability to assess CT scans concurrently with the PET scans. Another limitation was that the histological specimens could only be collected after the CT scan (50 min after injury). However, the collected data represent a very early phase of pulmonary injury.

Acid aspiration was performed as described in previous studies^{7,12,13,17–20}  in which the amount of instilled HCl varied (0.1–2 ml/kg). The amount of acid administered in the current study was in the lower range of effective doses, and this low volume is associated with a low mortality rate in rats.²¹  By administering HCl directly via a tracheal tube, aspiration injury can be controlled to affect only a certain part of the lung,²²  although severity is dependent on the pH and the amount of acid.¹⁷  This model has a definite, rather than diffuse, point of origin, which enables the researcher to study localized areas of injury. The insertion of the instillation catheter to a maximum depth makes it likely that the damage does not extend to the other side of the lung and that it affects only the basal and middle regions.

Similar to the previous study, although acid aspiration induced hypoxemia and hypotension with acidosis, the heart rate remained unchanged.¹³ 

Most of the PBF/perfusion data that published regarding ALI or acute respiratory distress syndrome describe the regional blood flow distribution after lung injuries such as surfactant depletion,^23,24  smoke inhalation,²⁵  and oleic acid^26,27  or endotoxin treatment.^28,29  PBF and perfusion data collected after acid aspiration have rarely been published³⁰  and to our knowledge, no blood flow data are currently available that concern the early phase (after 10 min) of acid-induced lung injury.

Both the control and the injury groups showed comparable entire lung ROI ratios between lung sides. However, the ROI_specific ratios between the groups diverge due to the increased PBF in the injured area (fig. 3). The left and right lungs demonstrated equal PBF distributions in uninjured lungs. The increased PBF for the injured side in the left-lung injury group may reflect the smaller left lung volume leading to a larger, relative acid dose compared with the right-lung injury group. This phenomenon was not related to the magnitude of density changes revealed by the CT scans, and no differences were found between the sides with regard to their proportions of higher density tissue. Therefore, it can be assumed that factors, other than the size of injured tissue, such as the local concentration of HCl, may play a role in PBF distribution. Other parameters may include varying damaged functional structures. The difference in PBF distribution may also originate from the relatively small numbers of animals in each injury group.

Centered on areas of increased PBF after injury, ROI_specific is associated with significantly increased PBF compared with the uninjured side. The increase in PBF within the region of injury also colocalized to regions of higher lung density in the CT scans. However, the PBF and CT values of the same region showed no correlation. This observation may be explained by the fact that the effect of the injury may not have extended throughout the entire ROI, which could have led to the underestimation of the calculated mean values. Another possibility is that the effect of the injury could have extended beyond the selected ROI, thus eluding detection. Because the PBF ratio varied from zero blood flow to maximal blood flow and the density in the same location varied from aerated to nonaerated, the maximal PBF in the injured region may have been accompanied by increases of varying magnitude in lung density. Increased maximal lung density may not predict maximal PBF (i.e., due to edema formation with resulting extravascular pressure in the lung parenchyma).

The PBF on the injured side was higher than that of the control animals for all selected transversal planes (apical, medial, and basal). There was a difference between the right- and left-side injury groups. These observations suggest that the injuries affected the selected ROIs, although it is likely that the injuries to the left lung evoked an increase of the contralateral PBF. We were unable to explain this pattern by the larger involvement of morphological or functional structures. There was no evidence of varying left-lung damage levels, either in differences in lung density or in terms of gas exchange and circulation when comparing the left- and right-lung injured rats.

The distribution of PBF is dependent on oxygen tension,³¹  and hypoxia was shown to cause similar regional changes in flow distribution in dogs and pigs,^32–34  with a shift of blood flow to the upper lung regions. A study in rats demonstrated that previous high-flow areas were hyperperfused during hypoxic exercise.³⁵  This type of hyperperfusion may have also occurred during local hypoxia in our study, resulting from blunted hypoxic pulmonary vasoconstriction after ALI.³⁶  During hypoxia, the endothelium releases relaxing factors that limit vasoconstriction in the pulmonary vascular bed.³⁷  Acid aspiration promotes local hypoxia and may attenuate hypoxic pulmonary vasoconstriction in injured regions, a finding also observed in sheep with pneumonia.³⁸  Hyperperfusion alone was shown not to be responsible for increased lung endothelial barrier protein permeability in sheep.³⁹  In contrast, hyperperfusion was associated with alveolar damage after single-lung ventilation in pigs.⁴⁰  The increased blood flow in areas of acid aspiration could intensify a loss of fluid through the damaged endothelium with subsequent local edema formation during the early stages of inflammation.

In a CT scan, an ideal method to assess lung damage,⁴¹  the impact of acid on lung tissue is represented by an increase in tissue density in the injured areas.^13,42,43  We considered the influence of microspheres on CT lung density to be negligible due to the human albumin structure of the particles. In our study, the regions of increased density within the right- and left-side injury groups varied but did not differ substantially. The involved tissue extended over nearly one fourth of the total lung and approximately one half of the injured lung. These injury dimensions had a notable impact on gas exchange and circulation.

Acid instillation creates a cellular injury to the alveolar epithelium and capillary endothelium,¹⁷  which has been characterized as the increased production of basement membrane components, increased numbers of intraalveolar erythrocytes, the recruitment of neutrophils, and fibrin formation within 4 h of injury.¹⁸  The cumulative DAD score showed no differences between apical regions but increased values for the lower ROIs of the injured side (fig. 6). Within the first hour after acid aspiration, histological abnormalities were localized to the areas of injury, and the differences observed with respect to the contralateral side were the result of inflammatory infiltration of neutrophils. Microatelectasis did not contribute to the cumulative increased DAD scores in basal regions (see figure, Supplemental Digital Content 3, https://links.lww.com/ALN/A963).

The histological investigations excluded the possibility of microsphere dropout from the alveolar capillaries. The samples did not demonstrate any accumulated microspheres due to microatelectasis or atelectasis, which implies that the PBF increases measured in this study were not the result of the consolidation of lung tissue.

The regional response of the lung to acid aspiration involves increased blood flow to the areas of injury and the deterioration of gas exchange during the early stages of the inflammatory response. The results of our study suggest that early treatments administered via the blood stream have the potential to reach injured lung regions effectively. This type of experimental approach may contribute to the development of further treatment strategies that could be applied promptly after lung aspiration.

The authors thank Regina Herrlich, Technician, for assistance with animal preparation and Jörg van den Hoff, Ph.D., Professor, both from the Institute of Radiopharmacy, Helmholtz-Zentrum Dresden-Rossendorf, Dresden, Germany; and Jonathan Yaqub, Medical Student, Technische Universität Dresden, Faculty of Medicine Carl Gustav Carus, Dresden, Germany, and Guido Musch, M.D., Associate Professor of Anesthesia, Department of Anesthesia, Critical Care, and Pain Medicine, Massachusetts General Hospital and Harvard Medical School, Boston, Massachusetts, for their insightful comments.