Objective　To study the injury factors, pathogenic process and clinical features of delay two-phase multiple organ dysfunction syndrome (MODS) in severe burned patients and to replicate a standardized animal model that would accurately imitate the clinical features of MODS.
　　Methods　Forty-five human patients with burn size larger than 30% total body surface area (TBSA) were analyzed. All of them underwent severe burn shock in early stage and sepsis in late stage. Thirty-two goats were randomly divided into three groups: 1) hemorrhagic shock (group H, n=6); 2) endotoxemia (group E, n=6); and 3) hemorrhagic shock plus endotoxemia (group M, n=20). Hemorrhagic shock was produced according to the method of Wigger (6.7 kPa for an hour, 1 kPa=7.5 mmHg). Endotoxin (E. coli O111 B4) was given via the portal vein 24 hours after the resuscitation of hemorrhagic shock, in a dose of 30 ng/kg/min for 5 consecutive days. During the observation period of 10 days, all animals were hemodynamically monitored, given standard metabolic support and due cardiac and pulmonary support according to human intensive care.
　　Results　All the patients showed burn shock at 1-3 days and hyperdynamic circulation, hypermetabolism and systemic inflammatory responses over two weeks post-injury. Thirteen cases were found to develop MODS according to the prevailing diagnostic criteria, and 10 of them died with a mortality of 77%. Eighteen animals died in group M with a mortality of 90%, 12 of the 18 developed MODS, with overall incidence of 60%. Most animals in group M showed changes similar to that observed in human cases. The experimentation proved that in the pathogenic process of MODS, there was a two-hit phenomenon in the dvelopment of the syndrome. To prevent the development of MODS, it therefore was imperative to blunt the first hit or the second hit, so that an excessive inflammatory response was alleviated. This postulation has been verified in the treatment of extensive burns. Two patients with burn extent reaching 100% TBSA survived with only mild acute respiratory distress syndrome (ARDS) and renal dysfunction after comprehensive treatment of burn shock, including adequate fluid resuscitation, drugs to remove oxygen free radicals, rapid restoration of pHi, and early extensive excision of burn eschars.
　　Conclusion　Both in human patients or animal experimentation, the typical delay two-phase MODS is shown to be produced by two successive insults in the forms of hypovolemic shock and sepsis. This postulation is helpful in formulating the prevention and treatment modality of MODS.

　

　　With remarkable improvement in our ability to prolong survival in patients with high lethal conditions, such as extensive burn injury or serious multitrauma, a new syndrome, termed multiple organ dysfunction syndrome, has emerged representing today the number one cause of death in such patients. Our treatment options of this sydrome are mainly supportive, and its pathogenic pathophysiology remains to be fully elucidated. To better understand the pathogenesis so that more promising therapeutic answers to the syndrome can be devised, experimentation to replicate an ideal animal model is mandatory. However, to develop such a model which faithfully simulates the pathophysiology and histopathology of multiple organ dysfunction syndrome (MODS) is as yet unsatisfactory, in spite that numerous experts have devoted to this problem. The present available models have not been universally accepted owing to infliction of a single injury factor in replicating the syndrome, short survival time of the animals, irrational diagnostic criteria, or lack of organ support in the process of reproduction.1-3　The present study is aimed at reproducing an animal model of better acceptance by introducing organ support and monitoring technics currently employed in the ICU into the process of reproduction. We are able to demonstrate that the injurious factors, pathophysiological changes, and clinical features in the animal model are all comparable with that of human patients.

METHODS

Patients data
　　As hypovolemic shock and sepsis are usually two distinct features of an extensive burn, burn patients were included in this study. A series of forty-five patients (men 34, women 11) suffering from burn injury exceeding 45% TBSA in extent were analyzed. Among them, 15 had burns of 30% 49% TBSA, 18 of 50%-69% TBSA, and 12 of 70%-100% TBSA. The mean age was 32.5±13.5 years, and mean full-thickness injury 40.5%±19.7% TBSA.

Examinations
　　Blood specimens were collected at 1, 2, 3, 7, 14 and 21 post-burn days (PBD) for blood counts, gas analysis, serum glutamic-pyruvic transaminase (SGPT), bilirubin, creatinine, urea nitrogen, and plasma amino acids profile. Blood culture was taken daily within the first week post-burn, and every other day after the first week.

Pathological study
　　Autopsy was done on 4 non-survivors, and specimens were taken from the lung, liver, heart, kidneys, and small intestine. The specimens were sectioned and stained for examination under both light and electron microscopy.

Grouping of animals and method of replication
　　In the process of reproduction, operative trauma, hemorrhagic shock and reperfusion injury constituted the “first hit”. The “second hit” consisted of introduction of endotoxin (LPS) into portal circulation. Throughout the experiment, organ supports were maintained.

Operation
　　Thirty-two goats were used. They were all male, aged one year, body weight 26.5±4.5 kg. After anesthetization with sodium pentothal 30 mg/kg intravenously, catheters were placed into the right jugular vein, right carotid artery and pulmonary artery. Tracheostomy was also performed. Through a midline abdominal incision, a catheter was passed into the superior mesenteric vein, and was pushed into the portal vein for measurement of portal blood flow (PVF). Gastrotomy and jejunotomy were done separately to insert tonometers for measurement of pHi.

Hemorrhagic shock and resuscitation
　　Twenty minutes after operation, hemorrhagic shock was induced according to the Wigger's method. Mean artery pressure (MAP) was maintained at 6.0-7.3 kPa for 1 hour. At the end of the shock, the shed blood, together with twice amount of Ringer's solution, was repidly reinfused through both the jugular vein and portal vein, restoring cardiac index and MAP to 80% of the preinjury levels.

Portal endotoxemia
　　Twenty-four hours after resuscitation, LPS of E. coli O111 B4 was infused 30 ng/kg/min through the portal vein for 5 days.

Monitoring and supporting organ functions
　　Respiration and circulation were monitored continuously. When cardiovascular or respiratory functions reached stage 1 (Table 1), artificial ventilation was started or dopamine was given to maintain MAP. Throughout the experiment, metabolic monitoring and metabolic support were instituted, with caloric rate twice of basic energy expend (BEE, actual measurement in goats was 85.5±3.9 kJ/kg/day), consisting of multiple amino acids, 10% fat emulsion and glucose by intravenous route. Daily protein amounted to 1.45 g/kg, with a calorie/nitrogen ratio of 180∶1. Animals received potassium 1.5 g/day and water 100-150 ml/kg daily.

Grouping of the animals
　　Animals were randomly divided into three groups: H group with hemorrhagic shock only (n=6), E group with intravenous LPS only (n=6), and M group with both hemorrhagic shock and LPS intravenous infusion (n=20).

Measurements
Cytokines and mediators
　　Plasma tumor necrosis factor (TNF), interleukin 6 (IL-6), interleukin 8 (IL-8) and the expression of transfer growth factor β (TGF-β) mRNA were determined.

Hemodynamics
　　Cardiac output, vascular pressures of systemic and pulmonary circulation and portal blood flow (PVF) were measured and cardiac index (CI), systemic vascular resistance index (SVRI), oxygen delivery (DO2), oxygen consumption (VO2), and oxygen extraction (O2ext) were calculated.

Table 1. Diagnostic criteria and scoring of MODS in animals

　Organ Dysfunction
Stage 1 (score 1) Stage 2 (score 2) Stage 3 (score 3)
Pulmonary PaO2 <8 kPa or 250 <PaO2/FiO2 ≤300 PaO2/FiO2 ≤250, PEEP ≤8 cm H2O, PaO2/FiO2 ≤250, PEEP >10 cm H2O,
　 　 　FiO2 ≤0.5 　FiO2 >0.5
GI Distension, 7.0<pHi≤7.1, L/M↑ Absence of peristaltic sound, reflux of Paralytic ileus, bleeding stress ulcers
　 　>2×normal 　gastric content, pHi ≤7.0 　
Cardiovascular High CI, low SVRI, dopamine High CI, low SVRI, dopamine Low CI, low SVRI, heart rates
　 　≤100 μg/kg.min, MAP≥9.3 kPa 　>100 μg/kg.min, MAP≥9.3 kPa 　<60/min, no response to inotropic
Hepatic Bilirubin or SGPT, GOT, LDH Bilirubin or SGPT, GOT, LDH 　
　 　≥2×normal >3×normal 　
Renal Creatinine >2×normal Creatinine >3×norma l
Hematologic 　 PT and PTT↑ >25% or platelets Disseminated intravascular coagulation
　 　 ≤50 000 　
CNS 　 Unconsciousness with response to Coma with no response to pain
　 　 pain only 　
　　1 cmH2O=0.098 kPa. L/M: urinary lactulose/mannitol ratio; GI: gastrointestinal tract; CI: cardiac index; SVRI: systemic vascular resistance index;

　MAP: mean arterial pressure; SGPT:serum glutamic-pyruvic transaminase; GOT: glutamic oxaloacetic transaminase; LDH:lactic dehydrogenase.

Metabolic parameters
　　Blood sugar content, free amino acids, urinary 3-methlyhistidine (3-Meth) and blood lactate were assessed sequentially.

Organ functions
　　Specific enzymological parameters were assessed to determine the functional states of the heart, the liver and the kidneys; blood gases were determined for lung function, diamine oxidase (DAO), intestinal pHi, and urinary lactulose/mannitol ratio were determined for the evaluation of intestinal mucosa barrier function.

Bacteriology and endotoxin level
　　Blood culture, plasma endotoxin level, and quantitative bacterial culture of various organs were determined.

Morphological evaluation
　　Histopathological examination of various organs was done.

The diagnostic criteria and scoring of organ dysfunction of failure
　　According to the diagnostic criteria and scoring established by the authors, whenever more than two organs showed signs of dysfunction or failure either simultaneously or sequentially, the animals were diagnosed as having MODS or MOF. Diagnosis and scoring of MODS for patients were performed according to Marshall's method.4

Statistical methods
　　Data were expressed as ±s, and were statistically evaluated by the Chi square test, probability of less than 0.05 was considered significant.

RESULTS

Clinical course and staging
　　All the patients underwent severe shock. Large amounts of fluids were given to them for resuscitation, and hemodynamic parameters became relatively stable after resuscitation. Following a period of stability, systemic manifestations for systemic inflammatory response syndrome or sepsis appeared, consisting of hyperdynamic circulation, persistent hypermetabolism, usually after PBD 8, in 25 patients. In 8 of them, blood cultures were positive, constituting 35% of septic patients. They either developed MODS or recovered after intensive treatment. In these patients, four clinical stages were discernible: 1) stage of hypovolemic shock (1-3 days post-burn); 2) stage of relative stability (3-5 days post-burn); 3) sepsis stage (2-3 weeks post-burn); 4) stage of organ dysfunction or recovery.

Mortality and incidence of MODS
　　Thirteen out of 45 patients (28.9%) developed MODS. They consisted of 8 men and 5 women, with mean age of 33.5±21.3% years, and mean burn extent 73.5%±15.6% TBSA (Ⅲ° 54%±19.1%). Ten patients died with a mortality rate of 76.9%. The number of failing organs and their relationship with mortality are shown in Table 2. The lung and intestine were the most frequently failed organs, and renal failure accounted for the highest mortality. Table 3 shows the functional scoring of failed organs and their relationship with mortality. It was shown that with the increase in score, there was an increase in the degree of failure, as well as in mortality.

Table 2. Correlation between incidence of dysfunction of
different organs and mortality in burned patients

　 Lung Heart Kidney GI Liver Hematology
Incidence 10 　 8 　　8 　 6 　 2 　　　5
No.of deaths 8 6 7 3 1 3
Mortality (%) 80 75 88 50 50 60
Pathological findings
　　Lung: bleeding into alveoli; interstitial and alveolar edema; neutrophilic infiltration; microthrombi formation in

Table 3. Correlation between the organ dysfunction scoring
and mortality in burned patients

　 Lung Heart Kidney GI Liver Hematology
　 　1 2 　1 2 　1 2 1 2 1 2 1 2
Incidence 4 6 3 5 4 4 3 3 2 0 2 3
No.of deaths 3 5 2 4 3 4 1 2 1 0 1 2
Mortality (%) 75 83 66 80 75 100 33 66 50 0 50 66
the vessels. Heart: interstitial edema; focal necrosis of myocardial fibers. Kidneys: ischemia of glomeruli; interstitial hyperplasia in mesenteric region; dilatation of tubules with granular casts in lumen; bleeding in medulla. Liver: degeneration of hepatocytes; focal necrosis with cellular infiltration. Gastrointestinal tract: mucosal edema; patchy hemorrhage and inflammatory cell infiltration in submucosa.

Animal model
Clinical stages
　　M group animals manifested four clinical stages: hemorrhagic shock (beginning of operation to end of shock); resuscitation (end of shock to 24 h after injury); septic stage (infusion of endotoxin to 7 days after injury); and recovery (7-10 days after injury). Animals wihich survived 7 days usually would recover and survive. All surviving animals were sacrificed at the end of 10 days.

Mortality rate and incidence of MODS
　　90% of animals died in M group (Table 4), 50% in H group, and 33% in E group. 60% of animals developed MODS in M group (Table 4), and 17% in both H and E groups.

Bacteriology and endotoxin levels
　　Bacterial cultures were all negative for blood and various organs before the injury, while they became positive before death. Positive rates of bacterial culutre for organ and blood were 100% and 64%, respectively, in M group, and theywere significantly higher than that of the other two groups. All the microorganisms isolated were gram negative rods, of which E. coli constituted 80%. Plasma endotoxin levels were all significantly increased after injury in all animals (Fig.1). After the infusion of LPS, endotoxin level in the portal blood was significantly higher than that in the peripheral blood in M and E groups 5-7 days after injury. Plasma endotoxin levels were remarkably higher in M group than in the other two groups.

Table 4. Mortality and incidence of MODS in M group

Stages Mortality 　 Incidence
Shock n % 　 n %
Resuscitation 1 5 　 　 　
Sepsis (days) 2 10 　 　 　
　1-3 5 25 　　4 　 20
　4-5 6 30 4 20
　6-7 4 20 4 20
Total 18 90 12 60

Fig. 1. Changes in plasma levels of endotoxin in the portal vein. * P<0.05 vs baseline (preinjury); ☆ P<0.05 vs group E; # P<0.05 vs group H.

Changes in cytokines and systemic inflammatory response
　　The obvious elevation of TNF, IL-6 and IL-8 after shock and reperfusion injury is displayed in Fig.2. Infusion of LPS resulted in further increase in their levels, the magnitude of which was significantly greater than hemorrhagic shock. Meanwhile, expression of TGF-β mRNA became stronger in monocytes. Although TNF showed an elevation in E group, it was still significantly lower than that in M group, indicating that inflammatory cells were “primed” by the first hit resulting in more remarkable activation of inflammatory cells by the second hit. At the same time, animals showed fever (>40℃), rapid heart rate (>130/min), increased respiratory rate (>30/min), leucocytosis (>2×104/L), and neutrophilia, all indicating septic response. The level of TNF peaked at the late stage of sepsis, about ten-fold of the preinjury level. There were also signs of hypoxia, low urinary output, hypotension, elevation of blood lactate, indicating the onset of septic shock.

Fig. 2. Changes in plasma level of TNF, IL-6 and IL-8 in M group. * P<0.05, compared with preinjury.

Changes in hemodynamics
　　As shown in Figs. 3-5, the changes in hemodynamics could be divided into four stages, namely: 1) Stage of low cardiac output and high peripheral vascular resistance, which were the characteristics of hypovolemic shock. In this

Fig. 3. Mean arterial pressure (MAP) in animals in each group. * P<0.05 vs baseline (preinjury); ☆P<0.05 vs group E; # P<0.05 vs group H.

Fig. 4. Cardiac output index (CI) in animals in each group. * P<0.05 vs baseline (preinjury); ☆ P<0.05 vs group E; # P<0.05 vs group H.

Fig. 5. Systemic vascular resistance index (SVRI) in each group. * P<0.05 vs baseline (preinjury); ☆ P<0.05 vs group E; # P<0.05 vs group H.stage,MAP, CI and ventricular stroke work index were all depressed, while SVRI was elevated, PVF decreased, and both pulmonary and portal vascular pressure lowered. The animals were in the state of hypodynamic shock; 2) Stage of reperfusion, in which hemodynamic parameters were all stable; 3) Stage of hyperdynamic and low peripheral vascular resistance. This stage occurred at 1-5 days after the injury, showing progressive rise in CI, peaking on the 5th post-injury day, and being about 135% over the level at the end of fluid resuscitation. Meanwhile, SVRI showed progressive lowering up to the 8th post-injury day. Fluid replacement failed to maintain MAP. There was also lowering of portal venous pressure and flow, indicating the inadequacy of blood supply and oxygen delivery to the GI tract; 4) Stage of decompensated low output and low peripheral resistance. This was the terminal stage. Cardiovascular system failed to respond to supportive therapy; MAP, CI and SVRI rapidly lowered, and death ensued.

Metabolic alteration
　　The main manifestation was persistent hypermetabolism, which was most marked at septic stage. About 1-3 days after injury, VO2 and O2ext were remarkably elevated, indicating CI and DO2 were increased to meet the need of heightened metabolic rate. Three to five days after injury, plasma branch chain amino acids (BCAA) began to lower, together with a rise in aromatic amino acids (AAA) and a lowering of BCAA/AAA ratio, indicating decompensation of metabolic process. An elevation in the ratio of phenylalanine and tyrosine signified impairment of hepatic function. Urinary excretion of 3-Methy increased, indicating increased catabolic metabolism of muscular protein (Table 5). This was most marked in animals developing MODS. Five days after injury, DO2, VO2 and O2ext began to lower, accompanied by hyperglycemia and lacticemia, indicating an inadequate utilization of oxygen in tissue, resulting in anaerobic metabolism in the cells (Figs. 6-8). Low metabolic rate appeared before death of animals, indicating metabolic failure in the cells, and the process seemed to be irreversible.

Changes in organ functions
　　Changes in functions of main organs of M group are shown in Table 6. According to the diagnostic criteria and

Table 5. Changes in plasma amino acids, lactate, urinary 3-Meth in M group

　 Preshock End of shock Post-injury (days)
1 3 5
BCAA/AAA 3.25±0.33 3.51±0.81* 2.40±1.65 　2.09±0.71* 　2.07±2.80*
Phe/Tyr 1.23±0.11 1.28±0.13　 　1.54±0.18* 1.43±0.22 1.38±0.26
Lactate (μmnol) 0.9±2.9 18.3±2.6　 9.8±2.8 　12.0±6.5　 　43.6±11.3*
3-Meth (μmol/24 h*u) 　162±40　 　284±158　 273±177 359±156 451±17*
　　* P<0.05, compared with preinjury. BCAA: branched-chain amino acids; Phe/Tyr: phenylalanine/tyrosine; 3-Meth: 3-methylhistidine; AAA: aromatic
　　amino acids.

Fig. 6. Systemic oxygen delivery (DO2) in animals in each group. * P<0.05 vs baseline (preinjury); ☆ P<0.05 vs group E; # P<0.05 vs group H.

Fig. 7. Systemic oxygen consumption (VO2) in animals in each group. * P<0.05 vs baseline (preinjury)； ☆ P<0.05 vs group E; # P<0.05 vs group H.

Fig. 8. Systemic oxygen extraction (O2ext) in animals in each group. * P<0.05 vs baseline (preinjury); # P<0.05 vs group H.

scoring, impairment of the intestinal mucosa barrier function appeared earliest and most frequent during shock and resuscitation period. Heart and lung dysfunction usually occurred 1-3 days after injury, while dysfunction of the liver, kidney, and coagulation occurred at 3-7 days after injury. It was also found that death rate was increased with the increase in scoring (Table 7).

Pathological changes
　　They were most marked in M group. The main findings in the lung were interstitial hemorrhage, alveolar edema and formation of hyaline membrane in the alveoli. The pathological changes in the kidneys consisted of acute inflammatory necrosis of renal tubules and the presence of numerous tubular casts. In the intestinal mucosa and liver, there were patchy hemorrhage, necrosis and infiltration of inflammatory

Table 6. Changes in organ functions in animals in M group

　 Preinjury End of shock Post-injury (days)
1 3 5 7
PaO2 (kPa) 　13.3±1.5　 　12.1±2.8　 12.0±2.2　 10.4±1.4* 9.2±1.3* 　9.9±0.5*
GPT (IU/L) 　23±6　 29±11 65±45* 　77±45* 59±29* 257±252
Cr (μmol/L) 80±17 99±35 100±43　 　128±28*　 129±77* 243±57*
LDH1/LDH2 3.1±1.5 3.7±1.9 4.8±1.8* 8.0±2.1* 7.9±1.9* 4.0±2.2
DAO (u/ml) 0.9±0.7 2.1±1.7 　1.2±0.2　 3.0±2.6* 4.7±1.2* 　
L/M <0.03 　0.4±0.3* 0.6±0.2* 1.7±0.6* 　 　
IpHi 7.30±0.06* 　7.05±0.06* 7.14±0.09* 7.05±0.07* 7.02±0.07* 　6.83±0.11*
　　* P<0.05, compared with preinjury. L/M: urinary lactulose/mannitol ratio; IpHi: intestinal pHi; Cr: creatinine; LDH1/LDH2: lactic
　dehydrogenase isoenzyme1/ lactic dehydrogenase isoenzyme2; DAO: diamine oxidase.

Table 7. Scoring of each organ function and mortality

　 Intestine Lung Heart Kidney Liver Hematology
1 2 3 1 2 3 1 2 3 1 2 1 1 2
No. of animals 　7 　3 　2 　3 　3 　3 　1 　4 　3 5 2 7 2 1
No. of deaths 1 2 2 1 2 3 0 2 3 3 2 3 1 1
Mortality (%) 14 67 100 33 67 100 0 50 100 　60　 100　 　43　 　50　 100　
cells. There was bleeding on endocardial surface and degeneration and necrosis of myocardium. The pathological changes in the E group were mainly congestion, edema, degeneration and cellular infiltration.

DISCUSSION

　　At present, MODS is still a topic which attracts enthusiastic discussion and research in the field of traumatic surgery and critical care. However, there is always a dispute in regard to its concept and diagnosis. In the 1990s, MODS is recognized as a disease state which is clearly related to strong insults (often with two hits) arising from shock and then infection or over inflammation, and it is probably an acute injury to various organs as a result of uncontrolled inflammation and systemic septic response.5　On the basis of this concept, the authors successfully reproduce an animal model which closely simulates the clinical picture of delay two-phase MODS that occurs in human patients. This animal model is different from previous ones in diagnostic criteria, injury factors, pathological pictures and clinical manifestations.

Diagnostic criteria
　　In this study, the authors proposed a new diagnostic basis for MODS: predisposing factors + systemic inflammatory response syndrome (SIRS) + multiorgan dysfunction. This diagnostic basis consists of three points: 1) The inducing factors should include a serious trauma and then infection, together with their respective supporting therapeutic measures; 2) It should reflect the characteristics of uncontrolled inflammatory response, hyperdynamic circulation, and persistent hypermetabolism; 3) There is dysfunction of at least two organs or systems.

　　The criteria which are enumerated here are basically different from previous ones, in which failure of more than two organs is the only criterion. It is obvious that the previous criterion dose not differentiate organ failure due to other causes, such as direct trauma to organ, from MODS we are defining. It also describes only the eventual outcome of the syndrome, which is the terminal event, but neglecting the characteristic pathophysiological features. Therefore, the previous diagnostic criterion fails to serve as a guidance in the research of the pathogenic mechanism of the syndrome, nor as an aid for early clinical diagnosis and treatment. On the contrary, the criteria that we are advocating lay emphasis on the entire clinical characteristics, i.e. predisposing factors, clinical features and developmental process, with organ failure as the final event.

Injurious factors
　　MODS is a syndrome which is the result of multiple factors. Clinical observation in burn injury reveals that delayed resuscitation of shock (prolonged ischemia-reperfusion injury), inhalation injury, burn wound infection and gutderived endotoxemia (sepsis of endogenous source) are the most frequent predisposing factors. In an effort to replicate MODS simulating clinical syndrome, the authors abandoned the former methods of animal models, which were usually produced with a single injury factor, representing either one of the predisposing factors or a certain facet of disease process or some of the clinical characteristics. Therefore, the success rate of replication is low, duplication is dubious, hence it has not been widely accepted by clinicians. Taking cognizance of the above mentioned short-comings, we reproduced the animal model using compounded injurious factors, such as hemorrhage shock following trauma, reperfusion injury, enterogenous septic response, etc., which are commonly found in the pathogenesis of MODS. In this way, we inflicted strong stress into successive episodes to the animals in the form of severe trauma together with hemorrhagic shock and endotoxemia by prolonged infusion of LPS into the portal circulation, rendering the animals to react fully to the injurious factors. Thus, we were successful to reproduce the entire course of decompensation of organ functions seen in typical clinical cases of delay two-phase MODS, which constituted about 70%-80% of all cases of the syndrome. Typically, the patients apparently rallies from initial insults, followed by a period of relative stability. Following a second insult, usually in the form of systemic infection or sepsis, sequential organ dysfunction and failure ensued, frequently ending in death.6

Pathogenetic process and clinical features
　　It was the opinion of Faist and Demling that multiple organ failure could be classified into two patterns, namely rapid single phase or primary multiple organ dysfunction, and delay two-phase multiple organ failure or secondary multiple organ dysfunction.6,7　In the former pattern, failure of organs is the result of direct injury or severe shock, and it occurs 1-3 days after injury, without evident manifestations of systemic inflammatory responses. We argue that in this type of multiple organ failure, the functions of the involved organ are either disrupted by direct external force, or as a result of sustained ischemia due to prolonged profound shock. Therefore, it should not be considered as true multiple organ failure in discussion.

　　The second type of organ failure, named as delay two-phase multiple organ failure, remains a baffling problem in the surgical field. It is unique that the organs which fail are not necessarily and directly injured or involved in the primary disease process, and there is usually a lag phase of days or weeks between the initial insult and the development of remote organ failure. There are prominent manifestations of systemic inflammatory response, indicating that organ failure is an external expression of interplay of multiple pro-inflammatory mediators. This is borne out by our clinical observation in burn patients. On the other hand, in our animal experimentation, through assessment of pro-inflammatory cytokines, determination of hemodynamics, analysis of sugar and protein metabolism, and evaluation of organ functions as well as histopathological examinations of organ, we prove that we are now possible to reprouce the typical delay two-phase MOF in an ovine model. The success in reproducing an animal model simulating faithfully the pathophysiology and histopathology of MODS further indicates that the present hypothesis of the pathogenetic mechanism of MODS is valid. Based on the pathophysiology that underlie this hypothesis, certain potential therapeutic approaches can be formulate. Furthermore, this ovine model can be used to test the validity of various therapeutic modalities before they are applied to human patients.

Intensive care for animals
　　Previously, efficient organ support after serious traumatic insult was neglected in the replication of animal models, so that it was difficult for the injured animals to survive shock or failure of a solitary organ. To assure the animals to survive primary insults, we institute a system of intensive care for the animals, so that the development of MODS runs its course without undue interruption. First of all, metabolic support is maintained throughout the observation period, and adequate nutrition is given by intravenous route, avoiding any interfering nutritional conditions encountered in previously reported experiments. For respiratory and cardiovascular support, the method and strength of support are more or less standardized to facilitate easy repetition. It is our belief that these supportive measures are pivotal in successfully establishing the animal model of MODS.

Treatment of MODS
　　Two-hit phenomenon in the pathogenesis of MODS implicates that the initial insult, such as hypoperfusion-reperfusion injury, primes the host so that on a second or subsequent insults, the host's response is greatly amplified, engendering dysfunction of various organs. Take an extensive deep burn as an example, hypovolemic shock and massive amount of dead skin tissue could most probably “prime” the patient, and subsequent infection and sepsis would result in an amplified systemic inflammatory response that damages the organs. Based on this concept, the best approach to prevent the development of MODS would be the removal of the deleterious effects of shock and dead tissue so that to shut down or blunt the inflammatory response and thereby restore a more normal physiologic state.

　　Two patients with 100% TBSA burn (Ⅲ° 40% and 56% TBSA, respectively), accompanied by moderate respiratory burn, were admitted on September 24, 1995. Under hemodynamic monitoring, the patients were resuscitated with large amount of fluids. Vitamins C and E and ginkgolide were administered to scavenge oxygen free radicals. Anisodamine was given to release mesenteric vasoconstriction so the pHi of gastric mucosa returned to normal within 48 hours. Thus both global and intestinal hypoxia was relieved. Twenty-two and 38 hours after burn injury, extensive excision of burn eschars was performed for the patients, with removal of 39% and 40% TBSA of burn eschars, respectively. The wounds were totally covered by allograft and microautografts. One patient was complicated by acute respiratory distress syndrome (ARDS), elevation of urea nitrogen to 15.9 mmol/L, creatinine 293 μmol/L, GPT 114 u/L, hypothermia (35.5℃), and hypernatremia (Na+ 155 mmol/L) on the PBD4. With mechanical ventilation support, appropriate antibiotics and repeated removal of necrotic tissue and skin grafting, the patient finally recovered two months post-burn. The other patient survived under similar therapeutic measures, and the clinical course was rather uneventful except a short period of septic symptoms.

　　The successful management of these two patients demonstrates well that timely treatment of all treatable harmful conditions is a promising therapeutic strategy to prevent or limit the development of MODS. In patients with extensive deep burn, early definitive primary and reoperative surgery, leading to the removal of necrotic tissue and successful coverage of the wounds, so that subsequent insult can be alleviated, are also of utmost importance.

Burn Institute, 304th Hospital, Beijing 100037, China (Hu S, Sheng ZY, Zhou BT, Guo ZR, Lu JY, Xue LB, Jin H, Su XQ, Sun SR, Li JY and Lü Y)
　　This research was supported by a grant from the Ministry of Health, General Logistic Department of PLA (the Eighth Five-Year Key Project), P.R.China.

REFERENCES

　　1. Wang P. An experimental animal model of multiple organ failure. Chin J Exp Surg 1987; 4:70.
　　2. Goris RJ. Multiple organ failure and sepsis without bacteria. Arch Surg 1986; 121:897.
　　3. Hu S, Sheng ZY, Xue LB. The experimental study of animal model of post-traumatic multiple organ failure. Chin J Plastic Surg Burns 1992; 8:2.
　　4. Marshall JC, Christou NV, Horn R, et al. The microbiology of multiple organ failure. Arch Surg 1988; 123(3):309.
　　5. Beal AL, Cerra FB. Multiple organ failure syndrome in the 1990s: systemic inflammatory response and organ dysfunction. JAMA 1994; 271:226.
　　6. Faist E, Baue AE, Dittmer H, et al. Multiple organ failure in polytrauma patients. J Trauma 1983; 23:775.
　　7. Demling R, Lalonde C, Saldinger P, et al. Multiple organ dysfunction in the surgical patient: pathophysiology, prevention and treatment.