Resuscitution. 22 Elsevier Scientific

( I99 1) 27-43 Publishers

Monkey hemorrhagic

27 Ireland

Ltd.

model of severe volume-controlled shock with resuscitation to outcome*

Gad Bar-Joseph, Peter Safar, Reisuke Saito”‘**, S. William Stezoski and Henry Alexander International Resuscitation Research Center (IRRC), Departments of Anesthesiology and Critical Care Medicine and ‘Pathology. and Presbyterian-University Hospital, University of Pittsburgh, 3434 Fifth Avenue, Pittsburgh, PA 15260 (U.S.A.) (Received

January

1 lth,

1991; accepted

February

lst, 1991)

Seventeen cynomolgus monkeys under N,O analgesia and sedation were subjected to severe volumccontrolled hemorrhagic shock (shed blood volume of 21 or 27 ml/kg). In I2 monkeys, resuscitation was started after increasing periods of hemorrhagic shock from 30 min to 5 h. In five additional monkeys. volume-controlled hemorrhage was modified at hemorrhagic shock 30 min to control MAP at 30 mmHg: resuscitation was started at hemorrhagic shock of 2 h. A clinically relevant resuscitation protocol consisted of a field phase from 0 to 6 h (lactated Ringer’s solution, spontaneous breathing), and a hospital intensive care phase from 6 h to 48 h (blood, lactated Ringer’s solution to mean arterial pressure (MAP) 2 JO mmHg, controlled ventilation, advanced life support). Fifteen of the I J monkeys survived. After outcome evaluation at 4 or J days, the eight monkeys with “moderate insult” had only transient functional impairment. Of the nine with “severe insult,” three showed signs of moderate transient non-oliguric renal failure. Eight of the I2 monkeys studied morphologically showed scattered liver cell damage. None of the monkeys developed pulmonary dysfunction or functional or morphologic evidence of cerebral damage. This study establishes a new hemorrhagic shock-resuscitation model simulating field-to-hospital life support. Severe hemorrhagic shock with MAP 30-40 mmHg for 90-120 min (without trauma or sepsis) can lead to complete functional recovery after transient malfunction of liver and kidneys.

Fluid resuscitation

-

Hemorrhage

-

Hypovolemia

-

Intensive

care -

Primates

-

Trauma

INTRODUCTION

The application of experimental results of resuscitation from hemorrhagic shock (HS) to clinical practice is problematic because the insults and therapies used in most existing animal models are unrealistic [ 1,2]. Dogs, which were used in most published studies, differ from primates or humans in their pathophysiologic responses to Address all correspondence and reprint requests IO: Gad Bar-Joseph, Pittsburgh, 3434 Fifth Avenue, Pittsburgh, PA 15260, U.S.A. *Presented ference of 1983 and **Current U.S.A.

IRRC,

University

of

at the annual meeting of the American Society of Anesthesiologists in 1982, the Annual Conthe Shock Society in 1982, and the World Congresses of Emergency and Disaster Medicine in 1985. address: Veterans Administration Hospital, Department of Pathology, Pittsburgh, PA 15213.

0300-9572/91/$03.50 Printed

M.D.,

and Published

0

1991 Elsevier Scientific

in Ireland

Publishers

lreland

Ltd.

28

bleeding [1,3--51. The Wiggers canine model of pressure-controlled HS [6], which has been widely used in the past, interferes with the spontaneous hemodynamic response to hemorrhage [l, 41. We previously developed a more realistic HS model in unanesthetized monkeys [7]. It included an initial volume-controlled blood loss that preserved the organism’s natural response for 30 min after end of hemorrhage. This was followed by pressurecontrolled arterial hypotension to standardize the severity of the insult. This volumepressure-controlled model was reproducible and, without fluid resuscitation, was lethal within several hours. Experimental design for testing the efficacy of therapies for HS should realistically simulate clinical scenarios. Sophisticated hospital-care measures, including mechanical ventilation and drug therapy, are often unavailable in the field. Titration of fluid resuscitation according to central venous pressure (CVP) or pulmonary wedge pressure measurements, commonly used in laboratory studies, is not possible in the field, where only noninvasive arterial pressure monitoring, pulse palpation, and perhaps measurement of urine output are available. Blood transfusion during the first hour, commonly used in laboratory studies [8-lo], is also unrealistic because it is rarely immediately available. It is surprising that late clinical complications of trauma and shock, such as acute renal failure or the adult respiratory distress syndrome (ARDS), were not seen in most animal models of HS. Oniy one report of an ARDS-like state was reported in monkeys resuscitated from HS [ 111. Although their pathogenesis is complex, these complications develop late, and short-term observations of 8-12 h, as practiced in most of the reported studies, may fail to detect them. A previous paper [7] described the pattern of dying. This second paper on the model [ 121 reports resuscitation and outcome to 7 days. It includes an experimental design for testing therapies in a realistic clinical scenario. This study had two objectives: (1) to develop a new monkey model of HS and resuscitation in two phases a field resuscitation phase followed by an in-hospital advanced intensive care phase; (2) to identify the point of irreversibility after the previously described insult [7], defined as that degree and duration of HS at which standard, optimal treatment would fail to prevent death or permanent, severe malfunction of heart, lungs, kidneys, liver, and brain. MATERIALS

AND METHODS

The animals were managed according to the guidelines of the American Physiologic Society. The HS insult used was described previously [7]. Briefly, male cynomolgus monkeys (A4ucaca fascicularis) weighing 3.3-5.6 kg, received endotracheal anesthesia with N,O/Oz, 75:25% and halothane, 0.5-l%, by intermittent positive-pressure ventilation (IPPV) during catheter insertion. Under sterile conditions, both femoral arteries were cannulated and a thermodilution pulmonary artery catheter was introduced. Electrocardiogram (ECG), systemic (SAP) and pulmonary artery pressures (PAP), end-tidal CO2 (ETCOz), electroencephalogram (EEG), and urinary output were monitored continuously. Halothane was discontinued about 30 min before bleeding, and spontaneous breathing was allowed to

Monkey no.

22 23 24

6 7 8

42 45

16b 17

21 27 27

27 21 21 21

I1 11

1 II 11 II 27 27

27 21 27 27

< 40 mmHg > 30 min) I 27 I 27 1 27

I I I

I I I I

no

yes yes

yes yes yes 120 120

37 120 120 I20

180 45 300

I06 Ill

50 118 90 98

120 48 38

Died Recovered

Recovered Recovered Recovered Recovered

Recovered Recovered Recovered

Recovered Recovered Recovered

0 29 21

90 75 120

no no no

no no no

Died Recovered Recovered Recovered

Recovered

19 26 0 I 5

90 60 60 90

4

RT (days)

Outcome

30

mmHg (min)

Final evaluation,

no no no no

(min)

to 30 mmHg

(ml/kg)

Total time of MAP < 40

no

Total HS time

MAP adjusted

Bleeding volume

< 40 mmHg < 30 min) I 27

Exp. series

“Died at RT 8 h from primary CNS failure. bDied at RT 33 h from primary cardiac failure.

28 31 37 39

I2 I3 14 15

Severe insult (MAP 9 25 IO 26 II 27

I8 I9 20 21

2” 3 4 5

Moderate insult (MAP I 17

Exp. no.

shock in monkeys. Retrospective stratification into moderate insult and severe insult groups. Series I for determination Table 1. Insult variables of hemorrhagic of irreversibility. Series II for standard therapy model. Total HS time refers to period starting at HS 0 until RT 0. and does not include time with shock during blood withdrawal. RT: resuscitation time.

30

return. During bleeding and HS, analgesia and sedation were maintained with N,O/Oz, 1525%. Hemorrhagic shock (Table I, Fig. I)

This report contains the results of two series of experiments in which HS and resuscitation protocols differed slightly. Bleeding was into phosphate-citratedextrose anticoagulant. Series I experiments (n = 12) were designed to determine the point of irreversibility of HS. The monkeys were arterially bled 21 ml/kg (n = 4) or 27 ml/kg (n = 8) at a constant rate over 20 min. The end of bleeding was considered HS time = 0. They were then allowed to respond naturally. After different HS times between 30 and 300 min (or when the MAP had dropped to < 30 mmHg for > 5 min) fluid therapy was initiated (at “resuscitation time” (RT) = 0). Series II experiments (n = 5) were designed to create a standard HS-resuscitation model as part of a study comparing various fluid therapies [ 131. The HS protocol was modified to prevent the large variability in hemodynamic responses to fixed volume bleeding alone [7]. The monkeys were arterially bled 27 ml/kg over 20 min. They were then allowed to respond naturally for the first 30 min. Starting at HS time 30 min, MAP was controlled at 30 mmHg, by withdrawal or transfusion of small increments of blood for an additional 90 min. After HS time 120 min, fluid resuscitation was started (RT = 0). Resuscitation (Fig. 1)

In both series, resuscitation was performed in two phases to simulate a clinical

-

Bpmt. breatMgFC v/O2

MAP

Rhger’s

n

pPVB_

75125%

Spent. breathitg

-

“Maintenance” + “Bolusea’

Ringer’s PxBV

Bbod TX

+

Hso

RT 0 30’ 2ll

RT6h

R-f 24 h

RT48h

RT 96h

Fig. 1. Diagram of volume-pressure-controlled hemorrhagic shock (HS) model, as used in series II experiments. The resuscitation period is divided into a field phase from resuscitation time (RT) 0 to RT 6 h, and a hospital intensive care phase from RT 6 h to 48 h. Outcome evaluation is to RT 4 days or RT 7 days. BV = blood volume. Blood TX = reinfusion of shed blood.

31

scenario: a field phase from RT 0 to 6 h, and a hospital intensive care phase from RT 6 to 48 h. In series I (n = 12), fluid resuscitation was started after increasing HS times (RT = 0). In the monkeys bled 21 ml/kg, resuscitation was started at HS 60 min (n = 2) or HS 90 min (n = 2); in those bled 27 ml/kg, resuscitation was started at HS 30, 75, 90, 120, 180, or 300 min (each n = 1), and in two other monkeys, resuscitation was started at HS 37 min or HS 45 min, when MAP decreased to < 30 mmHg for > 5 min. In series II (n = 5), fluid resuscitation was started at HS 120 min (RT = 0). In series I, Ringers solution with 5% dextrose, twice the shed blood volume, was infused over 30 min. At RT 30 min, this was followed by infusion of all the shed blood. Additional Ringers solution with 5% dextrose, 5 ml/kg per h was administered to RT 24 h. MAP of 2 70 mmHg was maintained with fluid boluses of 10 ml/kg. IPPV was used from RT 0 to RT 24 h, with pancuronium and diazepam as needed. Pat@ was maintained at 30-35 mmHg and Fro2 was adjusted to maintain Paoa > 80 mmHg. NaHCOs was given to correct a base deficit of > 6 mEq/l. In series II, an initial field phase to RT 6 h was followed by IPPV from RT 6 h to RT 24 h and intensive care to RT 48 h. The concept of field phase allowed no blood transfusion, no IPPV, and no drugs. At RT &-6 h, only one type of resuscitation fluid, Ringers solution with 5% dextrose, was infused at a rate according to MAP as described above. At RT 6 h, hospital admission was simulated and therapy included IPPV, NaHCO, i.v., as described for series I, including infusion of all the shed blood. The five monkeys of series 11are also the control group of a separate study that compared four different resuscitation fluids [ 131. At RT 24 h, all monkeys were weaned from IPPV, extubated, and monitored in the ICU until 48 h, when they were returned to their cages. Oral hydration was started in the ICU after extubation. Once in their cages the monkeys received water and food ad libitum. Evaluation

Arterial and mixed venous blood were sampled simultaneously and analyzed for Po2, Pco~, pH, Het, and 02 content (calculated from Hb O2 saturation in series I and measured directly by a Lex-O-Con analyzer in series 11). Cardiac output was measured by thermodilution and cardiac index (CI) was calculated. Renal function was assessed by determining the levels of blood urea nitrogen (BUN) and serum creatinine. Liver function tests included measurements of serum bilirubin, SGPT (AST) and SGOT (ALT). Pulmonary function was evaluated by Paoz, the alveolararterial oxygen tension gradient (PAo* - Pao,) and the respiratory index (RI) (PA02 - PaoZ/PaOz) [14]. Outcome at RT 4 days or 7 days was evaluated as overall performance category (OPC) (OPC 1 = normal, 2 = moderate disability, 3 = severe disability, 4 = coma) [15]. Results of the detailed neurologic and brain histologic evaluations are the subjects of another report. Statistical analysis included the unpaired t-test for group comparisons and the paired t-test for comparing variables within the same group at different times. After final neurologic evaluation the animals were reanesthetized and euthanized by infusion of 2% paraformaldehyde into the left cardiac ventricle [15,16]. A com-

32

MAP

mmHg

A 140

120

100

80

60

40,

20.

0’

I

I

’ EL

HS 0

HS 30’

HS 60

FIT 0

RT 30

RT lh

RT 4h

RT 12h

RT 24h

RT 48h

Cardiac Index ml/kg per mln WC

B

2oa

100

0 EL

HS 0

HS 60’

RT 0

RT lh

RT 4h

RT 12h

MAP (A) and cardiac index (B) at baseline (BL), at end of bleeding between RT 0 and RT 48 h. Moderate insult group (open circles) = HS with min; and severe insult group (closed circles) = HS with MAP < 40 mmHg > and cardiac index were significantly lower in the severe than moderate insult h, MAP and cardiac index in the severe insult group were significantly lower * = P < 0.05and ** = P < 0.01are for comparing moderate with severe # = P < 0.05 and ## = P < 0.01are for comparing with BL values. Fig. 2.

RT 24h

RT 48h

(HS = 0), during HS, and MAP c 40 mmHg < 30 30 min. During HS, MAP group. During RT 1-24 than baseline. Difference, insult groups. Difference,

33

plete necropsy was performed, and samples of all vital organs were obtained for histologic evaluation by light microscopy. Paraffin-embedded sections were stained with hematoxylin-eosin. RESULTS

Bleeding The response to hemorrhage up to the start of fluid resuscitation was reported previously [7] (Fig. 2). Briefly, in series I (n = 12), all monkeys showed a typical biphasic response. During bleeding, MAP progressively decreased to about 2CL-30 mmHg and then transiently increased to peaks of 54 f 17 mmHg. The degree and duration of this “self-resuscitation attempt” was quite unpredictable. While some animals had only minor, brief increases of MAP, in others MAP spontaneously recovered to almost baseline values for several hours. Cardiac index followed this biphasic course of MAP (Fig. 2). In series II, there was a similar biphasic natural response to blood loss, with attempted self-resuscitation in the first 30 min, but thereafter the variability in hemodynamic responses was eliminated by controlling MAP at 30 mmHg from HS time 30 min to HS time 120 min. All other hemodynamic variables followed a similar course. Control of MAP at 30 mmHg succeeded with titrated reinfusion or withdrawal of small increments of blood. The level and duration of hypotension determines the severity of insult. In series

HS 0

HS 60’

HS 120’

RT lh

RT 4h

Fig. 3. Arterial serum lactate levels (mean * S.D.) at baseline (BL), at end of hemorrhage (HS = 0 min), during HS, and at RT 1 h and 4 h. Moderate insult group (open bars) = HS with MAP < 40 mmHg < 30 min; and severe insult group (solid bars) = HS with MAP < 40 mmHg > 30 min. Difference, * = P < 0.05 and ** = P < 0.01 are for comparing moderate with severe insult groups. Lactacidemia significant in the severe insult group,,not in the moderate insult group.

34

I, MAP was not controlled. With increasing duration of HS and severity of hypotension, permanent damage to vital organs was expected. Consequently, all 17 monkeys were retrospectively stratified into moderate vs. severe insult groups (Table I, Fig. 2). The animals were assigned to the moderate insult group when MAP was ~40 mmHg for ~30 min; hypotension in this group actually lasted O-29 min (n = 8). They were assigned to the severe insult group when MAP was < 40 mmHg for > 30 min; hypotension in this group actually lasted 38-120 min (n = 9). The validity of this stratification was confirmed by serum lactate levels, which were higher in the severe insult group [ 171 (Fig. 3). Serum lactate levels were not significantly different between these two groups at baseline and at the end of bleeding (HS time = 0); but at HS time = 1 h, lactate levels were three times higher in the severe insult group (15.5 f 5.7 mmol/l) compared with the moderate insult group (5.3 f 3.2 mmol/l) (P c 0.01). This significant difference persisted during the first 4 h of the resuscitation phase, although lactate levels gradually decreased in both groups. Resuscitation

Fifteen of the 17 monkeys survived the 4 days or 7 days observation period. Functionally, all survivors had no neurologic impairment and had grossly normal behavior (OPC 1). One nonsurvivor had HS of 90 min, was in the moderate insult group, and died of primary CNS failure. It developed slightly depressed EEG activity shortly after start of fluid resuscitation, and required additional fluid boluses to maintain MAP. At RT 7.5 h the EEG suddenly became isoelectric, hypotension did not respond to additional fluid infusion, and cardiac arrest occurred at RT 8 h. Necropsy findings were negative. The other nonsurvivor had HS of 120 min, was in the severe insult group, and died of primary cardiac failure. At RT 15 min, there was a short period of bradycardia with MAP maintained, but then ECG showed prolonged PR intervals and ST segments, which persisted after heart rate normalized. At RT 8 h hypotension developed, was resistant to fluid therapy, and led to cardiac arrest at RT 33 h. There were unusual polyuria, increases in serum creatinine level to 2.2 mg/dl at RT 12 and 24 h, and a decrease in Pa@ before arrest. Necropsy showed mild interstitial edema of the heart and vacuolization of renal tubular epithelial cells. Cardiovascular function

In all 17 monkeys, the response to fluid therapy was prompt. MAP increased almost to control values during the initial infusion of lactated Ringer’s solution (Fig. 2). At RT 1 h, after the shed blood was reinfused, both MAP (Fig. 2A) and cardiac index (Fig. 2B) stabilized at control values. At RT 48 h, MAP tended to decrease, but was not significantly lower than control values. Cardiac index was reduced markedly during HS in the severe insult group, and reduced less in the moderate insult group at HS 60 min (P < 0.01) and HS 120 min (P < 0.05). During fluid resuscitation, cardiac index increased to control values at RT 1 h in both groups. Cardiac index then gradually decreased during the first 24 h to 62-76X of baseline in the severe insult group, but not in the moderate insult group (P < 0.05). At RT 48 h, cardiac index had returned to control levels in both groups (Fig. 2B).

35

Pulmonary function Except for the two nonsurvivors, all monkeys could be weaned from IPPV and were extubated according to protocol. During HS, all monkeys spontaneously hyperventilated. They reduced Pace, at the end of bleeding (HS = 0 min) to a lowest mean value of 22 f 6 mmHg. Pacoz then increased to 27 f 5 mmHg at HS 60 min and to 29 f 15 mmHg at 9@-120 min. After fluid resuscitation, still with spontaneous breathing, PaCo, remained at 3-O mmHg until start of IPPV at RT 6 h, when Pace;! was controlled. After discontinuance of IPPV and extubation, there was no dyspnea nor any sign of pulmonary edema or ARDS. Throughout the 30

A

25 20 15

10 5, -

a I-

l-

II BL

2.5 -

RT 4 h

RT 24 h

RT 48 h

B

2.0 -

1.5 -

1.0

0.5

0.0 i BL

RT 4 h

RT 24 h

RT 48 h

Fig. 4. Renal function variables at baseline (BL) and after HS at RT 4 h, 24 h, and 48 h. Mean & SD. Blood urea nitrogen (BUN) (A) and serum creatinine (B) after HS higher in severe than moderate insult group (*P < 0.05, **P < 0.01); and in severe insult group elevated above baseline values (#. P < 0.05, ##, P < 0.01).

36

hospital ICU phase, Pao, was easily maintained at >80 mmHg with FIOZ 0.25-0.33. Calculated PAO* - Pao2 remained < 50 mmHg in almost all measurements. There was no difference in PAO* - Pao, or in the respiratory index between the moderate and severe insult groups. The Pao, also remained constant during the hospital ICU phase. Renal function (Fig. 4)

Urine output ceased during HS in all monkeys, and recovered at RT 30-60 min following fluid resuscitation. Mean urinary output during the field phase of resuscitation was 54 ml/h, with no difference between severe and moderate insult groups. None of the monkeys showed signs of anuria or oliguria after start of resuscitation. BUN (Fig. 4A) was twice as high in the severe insult group than in the moderate insult group (P < 0.005) at both RT 24 and 48 h, but was still at the upper limit of normal values. Serum creatinine (Fig. 4B) was significantly increased from baseline values in the severe insult group, from about 1.0 mg/dl to about 1.6 mg/dl at RT 4, 12, and 24 h (P < 0.05). Creatinine was higher in the severe rather than the moderate insult group (P < 0.05). At necropsy, three of eight monkeys in the severe insult group had marked renal tubular dilatation. These three animals had the highest BUN levels (21-27 mg/dl) and serum creatinine levels (1.63.0 mg/dl) at 24 and 48 h. On day 7, one of these three monkeys had died and the other two had creatinine levels of 1.5 and 3.1 mg/dl. Urine flow in these animals continued throughout the 48 h hospital ICU phase. Liver function (Fig. 5)

Total and direct serum bilirubin levels remained normal in all monkeys throughout HS and resuscitation. Following HS, SGOT (ALT) (Fig. 5A) and SGPT (AST) (Fig. 5B) increased significantly. At RT 48 h, SGOT (ALT) was increased seven-fold in the moderate insult group, and 25-fold in the severe insult group, compared with baseline control values (both P < 0.01). Throughout the ICU phase, SGOT (ALT) was three to live times higher in the severe insult group compared with the moderate insult group (P < 0.05). SGPT (AST) increased only insignificantly in the moderate insult group, while in the severe insult group it increased significantly compared with controls (P < 0.05), with a peak value of 290 U/ml at RT 24 h. There was no correlation between SGOT (ALT) and SGPT (AST) values and morphologic changes in the liver. Morphologic changes

Necropsy was performed on 12 of the 17 monkeys (Table II). The changes can be divided into two patterns. The first includes tissue alterations directly related to hemorrhagic shock, i.e. ischemia and blood coagulation (marked with * in Table II). The second includes tissue alterations related to complications during the post-HS recovery phase, such as infectious processes. The heart showed focal myocardial necrosis in two monkeys. One monkey shows septic thromboemboli involving the heart and kidneys. An additional monkey also had septic thromboemboli in the heart.

37

BL

BL

RT 4 h

RT 4 h

RT 24 h

RT 24 h

RT 48 h

RT 48 h

Fig. 5. Liver function variables at baseline (BL) and after HS at RT 4 h, 24 h, and 48 h. Mean f S.D. ALT @GOT) (A) and AST (SGPT) (B) after HS in moderate insult group (open bars) vs. severe insult group (solid bars) (*P < 0.05) and compared with BL values (#, P < 0.05, ##, P < 0.01).

The lungs showed neither pulmonary edema, nor ARDS-like changes. Four animals had pulmonary thromboemboli, and three had pneumonia. The kidneys showed normal glomeruli in all specimens. Three of eight severely insulted animals had tubular dilatation. There was acute pyelonephritis in four monkeys. The gastrointestinal tract grossly showed no bleeding or mucosal ulceration. The stomach was microscopically normal in all animals. Only three monkeys showed small areas of mucosal necrosis, with or without thromboses, in the small intestine or colon. The spleen was congested in most animals, but necrotic areas were seen in only two.

38 Table II.

Necropsy

Heart Myocardial

results.

necrosis

Number

of animals

(focal)”

Septic thromboemboli Lung Thromboemboli”

with marked

morphologic

changes.

Moderate insult group

Severe insult group

(n = 4 autopsies)

(n = 8 autopsies)

2

0

I

I

3

I

2

I

3

5

I

2

0 0

3 4

I

0

Stomach

o/4

018

Small bowel mucosa Focal necroses (mild)”

o/4

218

Large bowel mucosa Focal necroses (mild)”

214

II8

l/4

I18

Focal pneumonia Liver Fatty

metamorphosis”

Centrilobular Kidney Tubular

necrosis”

dilatation

Interstitial inflammation or pyelonephritis Septic necroses

Spleen Focal necroses” “Tissue alterations

directly

related

to hemorrhagic

shock.

The liver was grossly normal but microscopically slightly abnormal. Eight of the 12 monkeys had fatty degeneration and three centrilobular necroses. The brains were macroscopically normal. Minimal histologic changes in cerebral neurons caused by ischemia are described in another paper. DISCUSSION

This study established a novel monkey model of severe volume-controlled, pressure-adjusted HS and resuscitation by combining the following features: (1) use of primates; (2) avoidance of surgical anesthesia during HS, by using N20 for analgesia and sedation and spontaneous breathing of F,Oz 25% (similar to air); (3) avoidance of heparinization; (4) allowing an initial natural response to volumecontrolled hemorrhage; (5) extending the low-flow insult and controlling its duration

39

by subsequent pressure control; (6) using an HS insult of long duration (2-5 h); (7) simulation of fluid resuscitation in the field; (8) simulation of subsequent delayed advanced hospital ICU resuscitation; and (9) intensive care to outcome evaluation at 4 days or 7 days. The results include several unexpected findings: (1) severe HS with MAP of 30 mmHg for more than 90 min did not result in irreversible HS; (2) using standard fluid therapy and intensive care in this study, the monkeys achieved complete functional recovery after HS times of up to 5 h, even when MAP was 3040 mmHg for 2 h (Table I); (3) biochemical and morphologic observations in these severely insulted monkeys revealed no significant changes in the heart, lungs, or gastrointestinal tract; there was no evidence of ARDS; (4) in the liver and kidneys, there were only mild to moderate transient functional changes and mild scattered histologic changes that were compatible with functional recovery at 4-7 days. One unique feature of the model is the biphasic MAP response during the first 30 min after end of bleeding [7]. In contrast to canine models of pressure-controlled HS [6], which have been justifiably criticized [1,4], our volume-controlled, pressureadjusted model initially maintains the natural response to blood loss. With subsequent control of hypotension and prevention of cardiac arrest standardized duration and degree of HS was achieved. The use of primates, which have a response to bleeding similar to that of humans [ 1,3-51, makes this a particularly clinically relevant model. In our rat HS model we found the same biphasic MAP response [ 17a]. In dogs, the initial response to bleeding differs from the human response. Also, gastrointestinal hemorrhages and necroses develop after severe HS [18]. While the dog’s gastrointestinal tract is more sensitive to HS than that of the human, canine kidneys seem to be less sensitive than human kidneys. As the gastrointestinal pathology dominates the clinical picture, it may obviate the detection of other system’s failure. The severity of insult in our model may permit the detection of differences in the efficacy of various resuscitation methods. Our model simulates severe, temporary blood loss in the field, resulting in severe shock, with control of that hemorrhage occurring naturally or with the assistance of a rescuer. There was a 2-h delay in fluid resuscitation, representing the period during which a bystander would merely provide an open airway. Field resuscitation was then provided with Ringer’s solution i.v. for 6 h (simulating a response by a paramedic). This was followed by a hospital 1CU phase with additional IPPV and other measures from RT 6 h to RT 48 h. This scenario is relevant for severe trauma in rural areas, disasters, or wars where field resuscitation and evacuation are delayed. A different scenario in combat casualties is massive exsanguination leading rapidly to cardiac arrest [2,19,20]. In cases of exsanguinating hemorrhage in the field, the chances of survival until arrival at a medical facility are poor [2,2 1,221, but not hopeless [23-261. Results might be improved by shortening the time to primary and definitive treatment [25,26]. The 2-h delay in fluid resuscitation in our model is relevant also for recent wars; for example, in Vietnam the time from battlefield injury to hospital arrival was 2-3 h [27]. In mass disasters, a delay of 6-S h to evacuation and hospital admission is also realistic [25]. We consider early blood transfusion in animal models of traumatic-hemorrhagic shock [8-101 to be an unrealistic scenario. Our model allowed only one type of plasma or blood substitute

40

at HS 2 h, for the first 6 h of resuscitation. In this paper, we presented data of the control (blood) group of a larger study that compared several blood and plasma substitutes [ 131. Irreversibility of HS must be defined. Some define it as failure to recover without therapy [28,29]; others, as failure to recover from the shock state with retransfusion of shed blood but no other treatment [ 1,6,18,30]. With use of the Wiggers model [6], irreversibility has been defined as the need for reuptake of 15-25% of the initially shed blood in order to maintain the desired level of hypotension [1,4-6,18]. None of these definitions has a correlate in clinical situations. Acute irreversibility of HS in humans, as defined by unresponsiveness to massive fluid infusions and use of vasopressors, or as defined by disseminated intravascular coagulation with bleeding diatheses [31], is very rare [ 1,2]. More frequently, irreversibility in patients resuscitated from HS is manifested as single or multiple secondary irreversible vital organ failure - of kidneys [32-361, liver [37-41], lungs [42-501, or brain [51,52]. Renal or pulmonary failure probably develop in patients only in the presence of additional factors, such as sepsis or severe tissue trauma [42,43]. Lethal renal failure after HS in animal models has not been reported. Irreversible pulmonary failure following HS was reported in only one monkey study [ 111. We did not demonstrate “irreversibility” in terms of mortality or permanent cerebral, cardiac, renal, or hepatic failure. The two deaths were not due to irreversible HS, but to unexpected, and still unexplained, sudden cardiac or CNS deterioration. Thus, our results differ from those of Rutherford et al. [5], who reported a 75% mortality due to irreversible hypotension in pigtail monkeys bled according to the Wiggers technique. In Rutherford’s study and in our study, the initial HS was similar, but blood loss was slightly larger in our monkeys, and shock was more pronounced in five of our monkeys. However, Rutherford started therapy after the uptake of 15% of shed blood volume, which he considered to be the point of irreversibility, while we preserved the natural response to bleeding. We found potentially permanent multifocal lesions in kidneys and liver, which did not cause lethal dysfunction. We found no damage to lungs, heart, or gastrointestinal tract. Neurologic recovery was complete, with only minimal histologic changes. Acute renal failure after traumatic shock and fluid resuscitation is associated with high mortality [32,33]. The incidence of acute renal failure was 1 in 200 severely injured casualties in the Korean War [34], and 1 in 600 in the Vietnam War [32]. The etiology of acute renal failure seems to be multifactorial, including tissue trauma, decreased renal blood flow, multiple blood transfusions with hemoglobinemia, nephrotoxic agents and sepsis [33,35]. In earlier wars, hypotension seemed to be a major factor in the early development of post-traumatic acute renal failure [33,34]. In Vietnam, as a result of rapid evacuation and early fluid resuscitation, renal failure developed several days after injury and was more likely the result of gastrointestinal trauma and sepsis [32]. Our severe insult group had higher and more prolonged elevations in BUN and serum creatinine levels than seen in the moderate insult group. We believe this is the first report of acute renal failure (although mild and transient) after pure HS and resuscitation in an animal model. The liver is sensitive to hypoxia and hypotension [37], which can result in cen-

41

trilobular necroses [38] and elevated liver enzyme levels [39]. In humans, liver dysfunction is very rare after HS [39] or nonhepatic trauma [40]. In our study, SGOT (ALT) and SGPT (AST) levels were much more elevated in the severe insult group, while serum bilirubin values remained normal. Two-thirds of our monkeys showed at necropsy fatty degeneration of liver cells, compatible with reversible injury; only three monkeys showed centrilobular necroses. AST and ALT levels did not correlate with the histologic damage. These enzymes are nonspecific and may not necessarily reflect liver cell necroses. Unfortunately, we do not have liver function studies at RT 4 days or 7 days, which might have correlated better with necropsy results. The inconclusive results obtained by us and others on permanent liver damage after severe, prolonged HS suggest that other factors, such as multiple blood transfusions with a high bilirubin load or sepsis are more responsible for liver dysfunction after trauma than is HS alone [41]. Progressive pulmonary consolidation (ARDS) was thought to be an expression of irreversible HS (“shock lung”) [42]. The lack of pulmonary damage in our study of severe, prolonged HS alone again suggests that ARDS develops only when tissue trauma, sepsis, fractures (fat emboli), brain trauma, or massive blood transfusions are present in addition to hypovolemic shock [29,43,45-501. HS in animal models has not resulted in ARDS [30,46,47,50]. The only exception is Rutherford’s study [ 1I], which showed a late onset (18-24 h) of pulmonary insufficiency after HS of 2 h in 38% of the monkeys. HS was reversed by fluid resuscitation and blood transfusion. Although in our study, also with HS of 2 h, MAP was lower (30 mmHg), our monkeys did not develop ARDS. Rutherford, however, used more fluid loading to maintain MAP near baseline (63 ml/kg plus 10 ml/kg per hour later), while we attempted only to maintain MAP L 70 mmHg (which required 54 ml/kg plus 5 ml/kg per hour later). We conclude that a clinically realistic monkey model of severe volume-controlled, pressure-adjusted hemorrhagic shock and field-hospital-type resuscitation is feasible. In primates, HS with MAP 3@--40 mmHg for up to 2 h, without tissue trauma or sepsis, can be survived beyond 4 days or 7 days without permanent, lethal vitalorgan failure. Such an insult causes only transient malfunction and nonlethal multifocal microscopic necroses in liver and kidneys. ACKNOWLEDGMENTS

Supported by the A.S. Laerdal Foundation and the Pennsylvania Department of Health. Lisa Cohn provided editorial assistance. Gale Foster helped prepare the manuscript. REFERENCES 1 2 3

B.W. Zweifach and A. Froneck, The interplay of central and peripheral factors in irreversible hemorrhagic shock, Prog. Cardiovasc. Dis., 18 (1975) 147-180. R.F. Bellamy, P.A. Maningas and B.A. Wenger, Current shock models and clinical correlations, Ann. Emerg. Med., 15 (1986) 1392-1395. G.S. Moss, An argument in favor of electrolyte solution for early resuscitation, Surg. Clin. North Am., 52 (1972) 3-17.

42

4 5 6 I 8 9 10

11

H. Swan, Experimental acute hemorrhage, Arch. Surg.. 91 (1965) 390-406. R.B. Rutherford and R.S. Trow, The pathophysiology of irreversible hemorrhagic shock in monkeys, J. Surg. Res., 14 (1973) 538-550. C.J. Wiggers, Physiology of Shock, Cambridge, MA, Harvard University Press, New York, 1950. G. Bar-Joseph. P. Safar, W.S. Stezoski, H. Alexander and G. Levine, New monkey model of severe volume controlled hemorrhagic shock, Resuscitation, 17 (1989) 1l-32. J. Dillon, L.J. Lynch, R. Myers and H.J. Butcher, The treatment of hemorrhagic shock, Surg., Gynecol. Obstet., 122 (1966) 967-978. A.V.S. Rayner, G.E. Lambert and R.L. F&on, Cardiac function during hemorrhagic shock and crystalloid resuscitation, J. Surg. Res., 24 (1978) 235-244. G.S. Moss, H.J. Proctor, L.D. Homer, C.M. Herman and B.D. Litt, A comparison of asanguineous fluids and whole blood in the treatment of hemorrhagic shock, Surg. Gynecol. Obstet., 129 (1969) 1247-1257. R.B. Rutherford, S. Arora, P.W. Fleming, T. Monaghan and D.H. Lowenstein, Delayed-onset pulmonary insufficiency in primates resuscitated from hemorrhagic shock, J. Trauma, 19 (1979) 422430.

12

G. Bar-Joseph, P. Safar, S.W. Stezoski and H. Alexander, Survival after severe hemorrhagic shock in monkeys: CNS, pulmonary, hepatic and renal outcome, Anesthesiology, 57 (1982) A97 (Abst). 13 G. Bar-Joseph, F. Jesch, P. Safar, J. Stremple and W.S. Stezoski, Stroma-free hemoglobin, hydroxyethyl starch, lactated Ringer’s, or blood, for severe prolonged hemorrhagic shock in monkeys, Crit. Care Med., I1 (1983) 219 (Abst). 14 M.S. Goldfarb, T.F. Ciurej, T.C. McAslan, W.J. Sacco, M.A. Weinstein and R.A. Cowley, Tracking respiratory therapy in the trauma patient, Am. J. Surg., 129 (1975) 255-258. 15 P. Safar, S.E. Gisvold, P. Vaagenes, H.H.L. Hendrickx, G. Bar-Joseph, N. Bircher, W. Stezoski and H. Alexander, Long-term animal models for the study of global brain ischemia (GBI). In: Protection of Tissues Against Hypoxia. Editors: A. Wauquier, M. Borgers and W.K. Amery. Elsevier, Amsterdam, 1982, pp. 147-170. 16 E.M. Nemoto. A.L. Bleyaert, S.W. Stezoski, J. Moossy, G.R. Rao and P. Safar, Global brain ischemia: a reproducible monkey model. Stroke, 8 (1977) 558-564. 17 R.A. Kriesberg, Lactate homeostasis and lactic acidosis, Ann. Intern. Med., 92 (1980) 227-237. I7a, D. Crippen, P. Safar, L. Porter and J. Zona, Improved survival of hemorrhagic shock with oxygen and hypothermia in rats, Resuscitation, 21 (1991) 271-281. 18 R.C. Lillehei, J.K. Longerbeam, J.H. Bloch and W.G. Manax, The nature of irreversible shock: experimental and clinical observations, Ann. Surg., 140 (1964) 682-710. 19 V.A. Negovsky, Resuscitation and artificial hypothermia (USSR), New York, Consultants Bureau. 1962. 20 B. Kirimli, S. Kampschulte and P. Safar, Resuscitation from cardiac arrest due to exsanguination, Surg. Gynecol. Obstet., 129 (1969) 89-97. 21 J.P. Smith, B.I. Bodai, A.S. Hill and C.F. Frey, Pre-hospital stabilization of critically injured trauma 22 23 24 25

26 27 28

patients: a failed concept, J. Trauma, 25 (1985) 65-70. C.R. Chudnofsky, S.C. Dronen, S.A. Syverud, J.R. Hedges and B.J. Zink, Early versus late fluid resuscitation. Lack of effect in porcine hemorrhagic shock, Ann. Emerg. Med., 18 (1989) 122-126. D.J. Torpey, Resuscitation and anesthetic management of casualties, JAMA, 202 (1967) 955-959. A. Trouwborst, B.K. Weber and D. Dufour, Medical statistics of battlefield casualties, Injury, 18 (1987) 9699. P. Safar, E.A. Pretto and N.G. Bircher, Resuscitation medicine including the management of severe trauma. In: Medicine for Disasters. Editors: P. Baskett and R. Weller, Butterworth and Co.. London, 1988, pp. 36-86. Y.L. Danon, E. Nili and E. Dolev, Primary treatment of battle casualties in the Lebanon War. 1982, Isr. J. Med. Sci., 20 (1984) 300-302. T.J. Whelan, W.E. Burkhalter and A. Gomez, The management of war wounds, Adv. Surg., 3 (1968) 227-23 1. W.A. Nelson and H. Swan, Hemorrhage: responses determining survival. Circ. Shock. l(1974) 213-285.

43

29 30 31

32 33 34

35 36

37

38 39 40 41

42 43 44 45 46 47 48 49 50 5I

52

R.F. Bellamy, D.C. Pedersen

and L.R. DeGuzman,

Organ blood flow and the cause of death follow-

ing massive hemorrhage, Circ. Shock, I4 (1984) 113-127. S.1. Kim, J.M. Desai and WC. Shoemaker, Sequence of cardiorespiratory alterations after gradual prolonged hemorrhage in conscious dogs, Am. J. Physiol., 216 (1969) 1044-1050. R.M. Hardaway, Pathology and pathophysiology of disseminated intravascular coagulation. In: Pathophysiology of Shock, Anoxia, and Ischemia. Editors: R.A. Cowley and B.F. Trump, and Wilkins, Baltimore and London, 1982, pp. 186197. A. Whelton and J.V. Donadio, Post-traumatic acute renal failure in Vietnam, a comparison Korean war experience, Johns Hopkins Med. J., 124 (1969) 95-105.

Williams with the

R.E. Lordon and J.R. Burton, Post-traumatic renal failure in military personnel in Southeast Asia, Am. J. Med., 53 (1972) 137-147. P.E. Teschan, R.S. Post, L.H. Smith, R.S. Abernathy, J.H. Davis, D.M. Gray, J.M. Howard, K.E. Johnson, E. Klopp, R.L. Mundy, M.P. O’Meara and B.F. Rush, Post-traumatic renal insufficiency in military casualties, Am. J. Med., I8 (1955) 172-185. H.H. Rasmussen and L.S. Ibels, Acute renal failure, Am. J. Med., 73 (1982) 21 I-218. K. Solez, The pathology and pathogenesis of human “acute tubular necrosis.” In: Acute Renal Failure, Correlations Between Morphology and Function. Editors: K. Solez and A. Whelton, Marcel Dekker, Inc., New York and Base], 1984, pp. 1742. R.A. Cowley, J.R. Hawkins, R.T. Jones and B.F. Trump, Pathology and pathophysiology of the liver. In: Pathophysiology of shock, anoxia, and ischemia. Editors: R.A. Cowley and B.F. Trump, Williams and Wilkins, Baltimore, MD, 1981, pp. 285-301. J.J. Lemasters, S. Ji and R.G. Thurman, Centrilobular injury following hypoxia in isolated, perfused rat liver, Science, 213 (1981) 661-663. R. Abizanda, J. Bergada, F.X. Valle, J. Lopez, E. Bosch and S. Garcia-Moris, The early detection of the effects of surgical shock on the viscera, Resuscitation, 9 (1981) 315-321. G. Nunes, F.W. Blaisdell and W. Margaretten, Mechanism of hepatic dysfunction following shock and trauma, Arch. Surg., 100 (1970) 546-556. H.R. Champion, R.T. Jones, B.F. Trump, R. Decker, S. Wilson, M. Stega, J. Nolan, R.A. Cowley and W. Gill, Post-traumatic hepatic dysfunction as a major etiology in post-traumatic jaundice, J. Trauma, 16 (1976) 65@--657. F.D. Moore, J.H. Lyons, E.C. Pierce, A.P. Morgan, P.A. Drinker, J.D. MacArthur and G.J. Dammin, Post-Traumatic Pulmonary Insufficiency, W.B. Saunders, Philadelphia, PA, 1969. P. Safar, A. Grenvik and J. Smith, Progressive pulmonary consolidation: review of cases and pathogenesis, J. Trauma, I2 (1972) 955-967. T.L. Petty and D.G. Ashbaugh, The adult respiratory distress syndrome, Chest, 60 (1971) 233-239. H.J. Proctor, T.V.N. Ballantine and N.D. Broussard, An analysis of pulmonary function following non-thoracic trauma, with recommendations for therapy, Ann. Surg., 127 (1970) 18&189. J.R. Meyers, J.S. Meyer and A.E. Baue, Does hemorrhagic shock damage the lung?, J. Trauma, I3 (1973) 509-519. B.C. Esrig and R.L. Fulton, Sepsis, resuscitated hemorrhagic shock and “shock lung,” Ann. Surg., 182 (1975) 218-227. P.E. Pepe, R.T. Potkin, D.H. Reus, L.D. Hudson and C.J. Carrico, Clinical predictors of the adult respiratory distress syndrome, Am. J. Surg., I44 (1982) 124-129. A.B. Montgomery, M.A. Stager, C.J. Carrico and L.D. Hudson, Causes and mortality in patients with the adult respiratory distress syndrome, Am. Rev. Respir. Dis., 132 (1985) 485489. J.F. Murray, M.A. Matthay, J.M. Lute and M.R. Flick, An expanded definition of the adult respiratory distress syndrome, Am. Rev. Respir. Dis., I38 (1988) 720-723. W.K. Dong, S.W. Bledsoe, D.Y. Eng, J.E. Heavner, CM. Shaw, T.F. Hornbein and J.L. Anderson, Profound arterial hypotension in dogs: brain electrical activity and organ integrity, Anesthesiology, 58 (1983) 61-71. A.G.B. Kovach and P. Sandor, Cerebral blood shock, Ann. Rev. Physiol., 38 (1976) 571-596.

flow and brain

function

during

hypotension

and