Pulmonary Edema Induced by Cerebral Hypoxic Insult in a Canine odel DAVID C. IRWIN, ANDREW VV. SUBUDHI, LISA KLOPP, DAVE PETERSON, ROBERT ROACH, AND ERIC J\10NNET
IRWIN
DC,
SUBUDHI
AW,
KLOPP
L,
PETERSON
D,
ROACH R, l\10NNET
E. Pulmonary edema induced by cerebral hypoxic insult in a canine
model. Aviat Space Environ Med 2008; 79:472-8. Introduction: The mechanisms of noncardiogenic pulmonary edema syndromes, such as neurogenic and high-altitude pulmonary edema, remain unclear even after years of study. Previous attempts to develop an animal model for these illnesses have used increased intracranial pressure or whole-body hypoxia. We hypothesized that a cerebral insult induced with a venous hypoxic blood infusion to the brain would trigger neurogenic pulmonary edema in a canine model. Methods: 'vVe measured indices of pulmonary edema. hemodynamics, nmepinephl'ine (,~E), and epinephrine values in anesthetized adult Walker hounds in which the brain was perfused for 2 h with either venous blood (venous perfused brain, VPB) or arterial blood (arterial perfused brain, APB) while maintaining normoxic pulmonary and systemic circulations. Normal cerebral perfusion was then reinstated for an additional 2-h period before euthanasia. Results: VPB animals showed a greater fall in arterial Po} and Sao, and higher peak plasma ,~E compared to APB. On necropsy, VPB animals had greater lung wet-ta-dry weight ratios compared to APB. Histological analvses revealed areas of marked alveolar infiltration of neutropl-lils and ~acrophages.J acute hemorrhage, congestion! and alveolar edema in the VPB animals. Discussion: This study supports the hypothesis that a cerebral insult from venous hypoxic blood can induce pulmonary edema. This method yields a promising approach to the study of noncardiogenic pulmonary edema syndromes. Keywords: cerebral" hypoxia, norepinephrine, pulmonary edema.
P
ULMONARY EDEMA IS a life threatening condition that affects thousands of visitors to high altitude each year (7). Models of studying pulmonary edema have largely been retrospective in nature, thus the etiology of the condition remains to be resolved. In order to advance the study of pulmonary edema, a prospective model must be developed. In 1966 Gerald Moss and colleagues proposed that cerebral hypoxia alone induced nuerogenic pulmonary edema (21). To test their hypothesis they developed a crude isolated cerebral hypoxia animal model in which venous "hypoxic" blood was shunted into an animal brain via the carotid artely, while pulmonmy and systemic Jlnormoxia" were maintained (22-24). Using this method they induced pulmonary edema in canines and other species (22). Although Moss et al. showed that the degree of hypoxia imposed by the venous perfusate was associated ,,\lith the development of pulmonary edema (4), they did not take into account other mediating or moderating factors, such as Pc0 2, pH, and metabolite concentration. Additionally, this landmark study has been subject to scrutiny since the description of the 472
surgical technique was vague, rendering replication difficult. Regardless, Moss's tedmical pioneering method may yet provide a useful model for a \vide spectrum of cerebral insults that subsequently induce pulmonary edema. These may include illnesses such as: 1) ischemic stroke or high altitude pulmonary edema in which pulmonary edema is associated vvith cerebral hypoxia (1); 2) hemorrhagic stroke in which blood Pc02 , pH, and metabolite concentrations affect the degree of cerebral insult (5); and 3) pulmonary edema formation after an epileptic event which may be induced from metabolite imbalance or abnormal l1erve stimulation (12). Therefore, we sought to further explore this model for the study of pulmonary edema subsequent to cerebral insult. Our primary goal was to validate and extend the technique reported by Moss et al. (22,24} Therefore, the present study was designed to address these questions: 1) could a level of cerebral "hypoxia" be achieved while systemic Jlnormoxia" was maintained; and 2) would venous "hypoxic'" blood induce pulmonary edema? We hypothesized that low cerebral P0 2 (brain hypoxia) would cause elevated pulmonary artery pressure and plasma catecholamine concentrations, resulting in a high pressure, high permeability pulmonary edema. Our results demonstrate that this model can be used to make the brain hypoxic while normal blood gases and oxygen content are maintained in the pulmonary and systemic vasculature. Furthermore, animals that From the Cardiovascular Pulmonary Research Group, University of Colorado Health Science Center, Denver, CO (n c. Irwin); the Department of Biomedical Sciences, College of Veterir,ary 'md Biomedical Sciences, Colorado State University, Fort Collins CO (L. Klopp, E. Mormet); the Department of Cardiac Surgery, Poudre Valley Hospital, Fort Collins, CO (D. Peterson); the Department of Biology, University of Colorado at Colorado Springs, CO (A. W. Subudhi); and the Altitude Research Center, University of Colorado Health Science Center, Denver, CO (R. Roach). This manuscript was received for review in October 2007. It was accepted for publication in February 2008. Address reprint reguests to: David C. Irwin, Ph.D., University of Colora_do Health Science Center (UCHSC), Caxdiova.scular Pulmonary Research Laboratory (CVP), School of Medicine, 4200 E 9 th Ave., Denver, CO 80262;
[email protected]. Reprint & Copyright @ by the Aerospace Medical Association, Alexandria, VA. DOl: 10.3357;' ASEM.2217.2008
Aviaiion, Space, and Environmental Medicine· Vol. 79, NO.5· May 2008
CEREBRAL HYPOXIA & PULMONARY EDEMA-IRWIN ET AL. received venous blood had greater lung wet-weight to dry-weight (LWW/LDW) ratios compared to control animals that received arterial perfusate, suggesting that cerebral hypoxia and/or CO2, pH, and metabolite changes induced pulmonary edema. METHODS
Animals All procedures were reviewed and approved by the Colorado State University animal care and use committee and 'vvere under the direct care of a veterinarian. Adult Walker hounds < 4 yr of age were enrolled in the study. To prevent hemodilution during the experiment, 2 wk prior to the surgical procedure 250 ml of blood were withdravvn from the jugular vein and stored at 4°C to "prime" the cardiopulmonary bypass pump the day of the surgery.
Experimental Design Six Walker hounds (25-30 kg) were randomly assigned to one of tv"o groups: 1) a venous blood perfused brain treatment group (VPB) that received hypoxic venous perfusate (from femoral and jugular veins) via the carotid artery (cerebral hypoxia); and 2) an arterial blood perfused brain group (APB) that received arterial perfusate (from the femoral artery) as above (cerebral normoxia). All animals underwent 2 h of cerebral perfusion followed by 2 h of "surveillance" as explained below under cerebral perfusion protocol. At the end of 4 h the animals were euthanized with sodium pentobarbital. Anesthetic protocol: Pre-anesthetic medications of fentanyl (2 f.Lg' kg-I; LV.) and medazolan (0.2 mg . kg-I; LV.) were administered 30 min prior to anesthesia. A.11esthesia was induced by thiopental (30 mg . kg-I) and maintained by inhalation of isoflourane and injections of fentanyl (0.01 mg' kg-I). Ventilation protocol: After anesthesia was induced, the animals were intubated and placed on a volume ventilator (10 breaths/ min; 35-40% O 2) to maintain arterial saturation (S302) greater than 95% during surgical isolation of the carotid arteries, femoral, and external jugular veins. Once venous and arterial isolations were achieved the animal \vas weaned to 21% O 2 and ventilation rate adjusted to ~8-1O breaths / min to maintain normal arterial blood gas tensions and oxygen saturation for the APB or VPB protocols, respectively. At the end of the 2 h of cerebral perfusion the animals were weaned from the respirator and allowed to spontaneously breathe room air for an additional 2 h ("surveillance" period) as described below in cerebral perfusion protocol. Surgical technique: Under sterile conditions, standard procedures were used to isolate the common carotid arteries, left external jugular vein, and left femoral vein and artery. Single femoral and jugular veins were cannulated with 14 Fr cannulas and joined into a single polyvinyl (Tygon, Saint-Gobain, Akron, OH) tube and connected to the inlet port of a cardiopulmonary bypass pump. These two veins were required to obtain the desired cerebral perfusate of S302 "'60%. A carotid artery Aviation, Space, and Environmental Medicine' Vol. 79, No.5' May 2008
was cannulated with an 8 Fr calUlula and connected to the pump outlet Initially, cerebral blood flow was maintained through the uncannulated carotid and vertebral arteries to prevent cerebral ischemia. The bypass pump was adjusted to deliver blood flow through the carotid artery to produce a perfusion pressure at 20 mmHg higher than the systolic blood pressure. As previously described (19), isolation of the cerebral circulation is achieved at this perfusion pressure. Briefly, retrograde blood flow is induced through the circle of Willis and thereby collateral arterial blood (e.g. vertebral and basilar arteries) is prevented from entering the brain. Cerebral perfusion protDcol: After normal carotid blood How was achieved from the cardiopulmonary bypass pump, femoral and jugular venous blood (VPB) (S,,02 ~60°1r») or femoral arterial blood (APB) for control conditions (S,,02 ~90%) was perfused through a single common carotid artery while the second carotid artery was ligated. Immediat'ely after ligation blood perfusio;1 was doubled through the remaining carotid artery, as determined with a flow probe (Transonic, Transonic Systems, Inc., Ithaca, NY) attached directly to the carotid artery to compensate for any loss of flow secondary to ligation. To limit contamination by oxygenated blood from the vertebral arteries, the perfusion pressure was adjusted over a period of ".15 min, to +20 mmHg above systemic systolic pressure. After 2 h of cerebral hypoxia, cerebral blood flow was returned to normal conditions by reinstating flow through the previous ligated carotid artery and discolUwcting the cardiopulmonary bypass pump. To reproduce the Moss's teclmical method, animals were allowed to spontaneously breathe room air for an additional 2 h with normal cerebral perfusion while they remained anesthetized before being humanely euthanized. According to Moss et al., it was in this acute postcerebral-insult period that pulmonary edema developed (20,22-24). Thus, the 2-h time period immediately after brain perfusion ("surveillance" phase) was a particular point of interest for study. Blood gas analyses: Blood samples were drawn from the venous or arterial perfusate entering the brain, jugular vein, dorsal pedal artery, and pulmonary artery for blood gas analyses every 10 min to quantify the degree of cerebral hypoxia under whole body normoxia. During the JJsurveillance" phase of the experiment, blood gases were drawn every 30 min for determination of lung ftmction. Hemodynamic measurement: A Swan-Ganz catheter was inserted into the pulmonary artery via the right jugular vein for detennination of pulmonaty artery and capillary wedge pressures. An 18-gauge over-the-needle catheter was placed in a dorsal pedal artery to measure mean arterial systemic blood pressures and heart rate (HR). The systemic and pulmonary artery catheters were connected to a transducer and continuously recorded on a Marquette 7000 monitor (Marquette, Milwaukee, WI). Cardiac output (CO) was computed from a standard thermodilution technique (4) using injections of 10 ml of iced saline, measured in triplicate for each time point (Monarch REF-l computer, Baxter Healthcare, Irvine, CA). 473
CEREBRAL H'{POXIA & PULMONARY EDEMA-IRWIN ET AL
Blood and tissue collection: Blood (5 ml) was withdrawn from the dorso-pedal arterial and jugular venous catheters at baseline (BL), 1, 2, 3, and 4 h of the experimental procedure and transferred into chilled vials containing EDTA (1 mg' mInI), aprotinin (500 KID . m! 1 of blood) and glutathione (1 mg' m! I), Plasma was separated by centrifugation (4°C; 14,000 x g; 10 min), frozen in liquid nitrogen, and stored at ·--80°C until assayed, A median sternotomy was performed and the heart and lungs were removed en bloc. The left caudal lobe was removed, weighed, and oven dried (65°C until stable vveight was achieved) for LWW/LDW ratios as indices of pulmonary edema. TIle remaining lung lobes were fixed under constant pressure (24 inches of H 20) in 10%, formalin (24 h), paraffin embedded, and sectioned to 4, !Lm for histological analyses. Histological analyses: Transverse sections from the hmg (three per lobe) were prepared and stained with hemotoxvlin and eosin. The slides "vere examined in a blinded fashion using the scale 0 = none, 1 = mild, 2 = moderate, and 3 = severe to evaluate the following categories: alveolar neutrophils, alveolar red blood cells, and interstitial and/ or alveolar edema. Scores were averaged across the entire lung for each animaL Catecholamine assay: As previously described (13), plasma concentrations of norepinephrine and epinephrine were analyzed by tandem high pressure liquid chromatography followed by mass spectrometry. Statistical analyses: Statistical analyses were performed with the JMP (version 5) statistical package (SAS, Cary, NC). LWW/LDW ratio, histological analyses, and peak plasma catecholamine concentrations were analyzed by non-parametric Wilcoxon Mann-Whitney tests. Hemodynamic measurements and blood gases were analyzed by ANOVA (type of brain perfusate) with repeated measurements (time), All results are expressed as means ± SENt Statistical significance was assumed at P < 0.05.
110 100
100
90
90
80 CIl
:t:
10
E 60 E
0 'Z
10
.a
60
50
40
c
SO
~
l
.~
50
"'" 0
40
W
Rl
30
30
20
20 A~ter~tJl p~lfuscdbraln
v~n('lUspvrftJ~wbl''ak!
,/lPB
VPB
Sys1~m~c
l....rteria~ blood
Fig. I. Blood gas analyses demonstrating venous blood perfusate (VPB) induced a cerebral "hypoxia" in the presence of systemic normoxia. Data represent mean .;- SEM of samples taken every 10 min during the 2-h cerebral perfusion. APB = arterial blood perfused brain animals; VPB = venous blood perfused brain animals. P < 0.001 vs. APB and systemic arterial blood. Significance relates to POlo Pco lo and S,02. Systemic arterial blood gases represent only the VPB animals demonstrating a degree of cerebral "hypoxia" in the presence of systemic normoxia.
the P0 2 and Sa02 were lower and Peo2 was higher in the venou "hypoxic" perfusate vs. systemic arterial blood (P < 0,001, Fig. 1). No diHerences in arterial blood pH were observed- between the ABP vs. VBP animals, Compared to BL values the P0 2 and pH decreased and Peo2 increased within 30 min of being weaned from the respirator in the APB animals (P < 0,001 BL vs. mean of "surveillance" period; Table I), and remained stable for the rest of the study period. However, Sa02 remained unchanged compared to BL values in the APB animals for the entire post-perfusion time period (Table I), Similar to APB animals, within 30 min of being weaned from the respirator, the P0 2 and pH decreased from BL values in the VPB animals and remained low (p = 0.003 BL vs. mean of "surveillance" period; Table I), and increased Peo2 values were noted (P = 0.004; BL vs. mean surveiUance" period; Table 1). In contrast to the APB animals, Sa02 decreased at 30 min post-perfusion compared to BL values and remained low for the rest of the study period in the VPB animals (P = 0.004 BL vs. mean f/
RESULTS
No differences were noted in Po2, Peo2, Sa02' or pH between the brain arterial perfusate and systemic arterial blood in APB animals (Fig. 1). However, in VPB animals,
TABLE I. ARTERIAL BLOOD GAS VALUES DURING THE 2-h POST-PERFUSION "SURVEillANCE" PERIOD. Post Cerebral Perfusion lime (h) Arterial Blood Sample Measurement P0 2 (mmHgJ Pc0 2 (mmHgl pH S;)02 (}~)
Saturation
Group APB VPB APB VPB APB VPB APB VPB
Baseline Arterial Blood 79
3 ,"JA 34,+,4 :"JA 7c41 .;- 0.02 ,"JA 90 :+: 3 :'JA -0-
0.5 (2 ..5) 67 51 59 47 7.19 7.26 ti7 73
+ 5 :+:4 .;6 :+: ,)r :+: 0.05 :+: 0.04 .;7 :+: 6
1 (::I) 65 46 58 54 720 7.24 92 71
5 :+:4 ~"±"~ 6 :+:4 .;- 0.1)', :+: 0.04 .;7 :+:5 -0-
1.5 (3.5) 65 43 60 50 720 7.22 87 67
.;-
:+:
5 r
,)
+ 6 :+: 6 :+: 005 :+: 0.05 .;7 :+: 7
2 (4,5)
65 42 59 53 7.20 7.23 H7 67
+ 5 :+:4 .;6 :+:4 :+: 0.05 :+: 0.0,1 .;7 :+: 7
Mean ± SEM 66", J')¥ 46 :+: 2* 59 + 3' 51 :+: 2"'+ 7.2.0 :+: 0.02* 7.24 :+: 0.02* HH :+:4 69 :+: 3*'
Data represent mean ,+, SEM of samples taken every' 30 min and pooled for either arterial-blood perfused group (APB) or venous-blood perfused group (VPB); numbers in parentheses () represent the total time of the expaimental protocol. + p.,; 0.002 vs. APB group; * P = 0.01 vs. baseline values. Baseline values were measured prior to the cerebral perfusion period and represent the mean:+: SEM of all six animals.
474
Aviaiion, Space, and Environmental Medicine· Vol. 79, No.5· May 2008
CEREBRAL HYPOXIA & PULMONARY EDEMA-IRWIN ET AL. iisurveillance" period). In the VPB animals, no changes in arterial blood pH values were noted at any time period in the 2-h surveillance period (Table I). V\'hen the means were calculated across all time points for the 2-h surveillance period and compared between the APB and VPB groups, P02, Pe0 2, and Sa02 were greater in the APB vs. VPB animals (P
