Postinjury Multiple Organ Failure: A Bimodal Phenomenon : Journal of Trauma and Acute Care Surgery (original) (raw)
Keywords: Multiple organ failure, Early multiple organ failure, Late Multiple organ failure, Postinjury systemic inflammatory response syndrome, Postinjury pneumonia, Dysfunctional, Inflammation.
"The connection between cause and effect has no beginning and can have no end." --Leo Tolstoy: War and Peace
Multiple organ failure (MOF) emerged as a new syndrome two decades ago as a result of better intensive care unit (ICU) technology. [1-3] Soon thereafter, it was recognized to be the leading cause of late postinjury death. [4] Throughout the 1980s, the care of the injured patient improved dramatically. [5] With implementation of optimal regionalized trauma systems, it is speculated that MOF is disappearing. [6] This has not been our experience, and apparently not that of other trauma systems worldwide. [7-15] Moreover, in the past decade, there have been no new, proven effective, interventions for postinjury MOF. [16-19] In part, we believe this academic stagnation is owing to inadequate characterization of the disease. [20] The existing epidemiologic data may be misleading for several reasons. [1-4,7-15,21-27] First, the often quoted studies from the early 1980s, tended to overrepresent penetrating wounds and are now outdated. Secondaly, many studies combine nontrauma with trauma patients. Third, the majority of studies were retrospective reviews where important clinical events could not be adequately documented. Finally, the earlier definitions of MOF were subjective and inconsistently applied across studies. Therefore, the purpose of this study was to prospectively analyze high-risk trauma patients admitted to a single regional trauma center to better define the epidemiology of MOF in the 1990s.
METHODS
Study Population
During a 4-year period ending December 1994, 457 patients with an Injury Severity Score (ISS) more than 15 and age more than 16 years who were admitted to the trauma ICU at Denver General Hospital and survived more than 48 hours, comprise the study population. Our approach to the injured patient remained constant during the study period. [5] All trauma patients were initially admitted to one of three general surgery teams. Each team consisted of five general surgery residents at following postgraduate year (PGY) levels: one PGY-5, one PGY-3, two PGY-2s (one specifically assigned to the trauma ICU) and one PGY-1. The care of these patients was directed by existing protocols and supervised by one of four general surgeons with expertise in trauma and critical care. Emergency department and operating room protocols were supervised by the Chief of Trauma (E.E.M.), trauma ICU protocols were supervised by the ICU Director (F.A.M.) and the ICU Clinical Specialist (J.B.H.), and the data base was managed by the Trauma Research Fellow (A.S.). All patients were prospectively followed until discharged from hospital or death.
Primary Outcome
The presence of adult respiratory distress syndrome (ARDS) and MOF were scored by standard scales. Our ARDS score (depicted in Table 1) is a modification of two commonly used scores [28,29] and our MOF score (depicted in Table 2) is a modification of our previously published score. [19,30] We have used these scores in other studies and consistently during this prospective study. Briefly, ARDS was evaluated daily by tracking five variables which were graded on a scale of zero to four (0 = normal, 4 = severely abnormal). This 20-point score was converted into a lung dysfunction score of 0 to 3. Likewise, heart, liver, and kidney were evaluated daily for dysfunction on a scale of zero to three (0 = normal, 3 = severe derangement). Individual organ failure was defined as an organ dysfunction score greater than or equal to 2 and MOF was defined as the sum of the individual organ dysfunction scores, simultaneously obtained after 48 hours of admission, greater than or equal to 4. We do not use organ dysfunction scores obtained in the first 48 hours to define organ failure because they may reflect reversible derangements induced by the inciting event or incomplete resuscitation. Laboratory data are obtained by routine patient care protocols. When a value required for daily organ dysfunction scoring was missing, the previous value was used. If no previous value was available, then the organ dysfunction score was 0. MOF was defined as "early" if it was present on hospital day 3 and "late" if it occurred after day 3.
Postinjury ARDS score.
Postinjury MOF score.
MOF Risk Factors
Based on previous studies, MOF risk factors, potential confounders, or effect modifiers were prospectively recorded. [7,20] These variables were collected under three headings: A. Host Factors, B. Tissue Injury Severity, and C. Clinical Indications of Shock. A. Host factors include age, sex and comorbid disease. Definitions of comorbid disease were derived from two sources. [31,32] Our modified definitions were: Heart disease: previous myocardial infarction or cardiac surgery, current use of cardiac medications including anti-arrhythmics, diuretics when prescribed for cardiac failure, inotropic agents, and medications for coronary artery disease. Treatment of hypertension without cardiac failure was not considered significant. Lung disease: asthma or chronic obstructive pulmonary disease requiring bronchodilators. Liver disease: admission bilirubin > 2 mg% (before major surgical procedure or massive transfusion), past episode of hepatic encephalopathy, or established cirrhosis. Kidney disease: chronic renal disease requiring dialysis; admission creatinine > 2 mg% or renal transplant. Diabetes mellitus: diabetes requiring medication. Malignancy: leukemia, multiple myeloma, non-Hodgkin's lymphoma, or solid tumor with metastasis. Immune dysfunction: AIDS or radiation, chemotherapy, or prednisone (0.3 mg/kg/day or equivalent dose of other steroids) within the last 6 months. Drug abuse: documented withdrawal syndrome related to ethanol or illicit drug. Neurologic Disease: chronic seizure disorder requiring medication and Other: any disease that does not fit the above criteria but was clinically relevant and required medication. B. Tissue injury severity included mechanism of injury, Glasgow Coma Scale (GCS), ISS, isolated major body region injuries, and predominant body region injuries. The latter two indices were derived from Abbreviated Injury Scores (AIS-1990). [33] Isolated major injury was defined as an AIS greater than or equal to 4 for the body region of interest (i.e., head, chest, abdomen, pelvis/extremity) with AISs of other body regions < 2. A predominant injury was defined as an AIS of greater than or equal to 4 for the body region of interest with AISs of other body regions less than or equal to 4. Control groups for comparison to isolated major injury or predominant injury excluded patients with an AIS greater than or equal to 4 of the body region of interest. C. Clinical indicators of shock included systolic blood pressure on admission to the emergency department (ED SBP), units of red blood cells (RBCs) transfused, worst value of arterial base deficit (Radiometer ABL 300, Copenhagen) and worst value of lactate (2 mL of heparinize arterial blood, YSI 2300 Glucose/L-Lactate Analyzer, Yellow Springs, Ohio).
These variables were collected in the following three time windows: (a) admission: data included ED SBP, and GCS; (b) 0 to 12 hours of admission: included units RBCs transfused, the worst value of arterial base deficit, and worst value of lactate; (3) 13 to 24 hours of admission: included the worst value of base deficit and lactate. Age, sex, comorbid disease, injury mechanism, and AIS of all injuries were determined as soon as the information became available.
Complications
Patients were monitored for the development of infectious and noninfectious complications. Infectious complications were categorized as major or minor. Major infections included pneumonia, empyema, lung abscess, abdominal/pelvic abscess, extensive wound infection, meningitis and "other" major infections. Pneumonia was diagnosed based on the following criteria: (A) chest-roentgenogram infiltrate persistent for more than 48 hours; (B) temperature > 38degreesC; (C) many polymorphonuclear cells on sputum Gram stain; (D) leukocytosis (WBC > 12,000/mm3) or leukopenia (WBC < 4000/mm3); (E) blood culture positive for the same pathogen recovered in sputum culture; (F) bronchoalveolar lavage quantitative culture with a pathogen growth > 103 colony-forming units/mL; and (G) histopathologic diagnosis (autopsy or open lung biopsy). Pneumonia was defined as one of the following combinations of the above described criteria: A + E; or A + at least two of B, C, D; or A + F + at least one of B, C, D or G. Pneumonia was excluded when there was clinical resolution without antimicrobial therapy or when an alternative diagnosis was definitely made clinically or by autopsy. Lung abscess was diagnosed based on clinical and radiologic evidence, while empyema or abdominal abscesses were defined as a purulent collection that required drainage. Major wound infections were those that required debridement. Meningitis was diagnosed by Center for Disease Control (CDC) criteria. [34] Other infections were classified as major if they were associated with septic shock (e.g., urosepsis). Minor infections were urinary tract infections, catheter-related infections, and wound infections not requiring debridement, sinusitis, and conjunctivitis. These were diagnosed using the CDC definitions. [35] Nonseptic complications were defined as noninfectious potentially adverse events that required treatment. These complications were recorded on the day the definition criqteria were met.
To determine the potential causal relationship between major infectious complications and MOF, we examined the temporal relationship between the onset of infection to serial MOF scoring. [35] We specifically compared the MOF score obtained on the day the complication was diagnosed and the MOF score obtained 48 and 72 hours later. Therefore, major infections could be classified in four categories: (1) "not related" because it was community acquired or occurred in the hospital 4 or more days before the onset of MOF; (2) potential "trigger" if the MOF score on the day of diagnosis was < 4 (i.e., no MOF) and rose greater than or equal to 3 points within 48 to 72 hours; (3) "worsening MOF" if MOF was present on the day of diagnosis (i.e., MOF score > 4) and rose greater than or equal to 3 points within 72 hours; or (4) potential "symptom" if the major infection occurred after MOF was present and was associated with a rise in the MOF score < 3.
Statistical Analysis
Univariate analysis was performed using2 with Yates correction or Fisher Exact test. Student's t test or the Welch alternate t test (for normal distributed populations with different variances) were applied for continuous variable. A p-value less than 0.05 was considered significant.
Risk factors for MOF were identified by multiple logistic regression analysis using "Forward" selection (Proc Logistic, SAS for Windows, version 6.0, SAS Institute, Inc. Cary, NC). Based on previous studies, the various risk factors and confounders were included into one of five models and for the continuous variables the following cutpoints were used: age more than 55 years, ISS greater than or equal to 25, GCS < 8, ED SBP < 90 mm Hg, 0 to 12 hours RBC transfusion > 6 units, 0 to 12 hours base deficit > 8 mEq/L, 13 to 24 hours base deficit > 8 mEq/L, 0 to 12 hours lactate > 2.5 mmol/L, and 13 to 24 hours lactate > 2.5 mmol/L. [7,20] The five models were derived using sequential sets of patient data collected over the first 24 hours and were as follows: (1) admission model: included the 441 study patients who had complete admission data (age, blunt or penetrating mechanism of injury, preexisting disease, ISS, GCS, ED/SBP); (2) 12-hour data with base deficit model: included 342 patients with all of the above admission data plus 0 to 12 hours RBC units transfused and 0 to 12 hours base deficit levels; (3) 12 hours with base deficit and lactate model: included 246 patients with all of the above data plus 0 to 12 hours lactate levels; (4) 24 hours data with base deficit model: included 274 patients with all of the above data plus 13 to 24 hours base deficit levels; and (5) 24 hours base deficit and lactate model: included 158 patients with all the above data plus 13 to 24 hours lactate levels. Interaction between variables was tested introducing interaction terms; however, no significant interactions were identified. Subsequently, using the same stratification, models were derived for "early MOF" and "late MOF." "Early MOF" models were obtained by excluding late MOF patients, while for "late MOF" models all early MOF patients were excluded. This provided a common control group (i.e., all non-MOF patients) for all models (i.e., all MOF, early MOF, and late MOF). The purpose of performing these various models was to identify consistent risk factors for "all MOF", "early MOF" and "late MOF" and to allow a comparison of relative importance by comparing adjusted odds ratios.
Survival function against time estimates of early MOF and late MOF were computed by the product-limit method. Kaplan Meier curves were obtained and compared by the Wilcoxon Test (Proc Lifetest, SAS for Windows, version 6.0, SAS Institute).
RESULTS
Of the 457 study patients, 353 (77%) were men, mean age was 36.0 +/- 0.7 years (see Figure 1 for age distribution) and 50 (11%) had documented comorbid disease. Injury mechanism was blunt in 330 (see Figure 2 for injury mechanism distribution). Mean GCS was 11.5 +/- 0.2 (range, 3 to 15), while mean ISS was 25.1 +/- 0.4 (see Figure 3 for ISS distribution). Eighty patients (18%) had an ED SBP < 90 mm Hg. Overall, the mean number of RBC units transfused during the first 12 hours was 5.2 +/- 0.4 units (range 0 to 60 units). Two hundred eighty patients (61%) required at least one emergent/urgent operation within the first 24 hours, which included 170 laparotomies, 59 craniotomies, 45 thoracotomies, 45 orthopedic procedures, and 23 other procedures. Additionally, 117 patients (26%) underwent one or more operations after 24 hours, which included 99 laparotomies, 96 orthopedic procedures, 34 craniotomies, 5 thoracotomies, and 36 other procedures.
Distribution of study patients by age.
Distribution of study patients by mechanism of injury.
Distribution of study patients by Injury Severity Score (ISS).
Seventy (15%) of the 457 study patients developed MOF. Host factors for patients with and without MOF are depicted in Table 3. While MOF patients were older, the incidence of preexisting disease was low and only the presence of heart disease was associated with MOF. Additionally, when all comorbid diseases were combined, there was no significant association with MOF. Tissue injury severity indicators for patients with and without MOF are depicted in Table 4. Interestingly, both isolated major and predominant head injuries had a lower incidence of MOF. Conversely, both isolated major and predominant abdominal injuries had a higher incidence of MOF. Finally, the summary of multiple injuries by the ISS was associated with MOF. Clinical indicators of shock in patients with and without MOF are depicted in Table 5. Shock indicators associated with MOF included 0 to 12 hour units of RBC transfusion, 0 to 12 hours base deficit, 13 to 24 hours base deficit, 0 to 12 hours lactate, and 13 to 24 hours lactate.
Host factors for patients with and without MOF.
Tissue injury severity for patients with and without MOF.
Clinical indicators of shock for patients with and without MOF.
Complications for patients with and without MOF are shown in Table 6. Not surprisingly, MOF patients suffered significantly more major infections, minor infections, and nonseptic complications. Additionally, MOF patients required more mechanical ventilator days (18.6 +/- 1.5 vs. 3.1 +/- 0.3, p < 0.001), longer ICU stays (25.6 +/- 2.2 vs. 8.7 +/- 0.5 days, p < 0.001). and had a higher mortality (25 of 70 patients (36%) vs. 13 of 387 patients (3%), p = 0.001). The mortality per number of organ failure was: single, 11% (7/62); two organs, 24% (7/30); three organs, 60% (6/10); and four organs, 62% (8/13).
Septic and nonseptic complications for patients with and without MOF.
The temporal distribution of the onset of MOF is shown in Figure 4. In 27 patients (39%), MOF occurred early (was present on day 3), while in the remaining 43 patients (61%) MOF presented late (more than 3 days). The pattern of organ failure at presentation of early MOF versus late MOF is shown in Table 7. At presentation, early and late MOF had a similar high incidence of lung failure, while early MOF had more heart failure, and late MOF had more hepatic failure. Independent risk factors for "all MOF," "early MOF," and "late MOF" were next identified by multiple logistic regression analyses using the five sets of patient data. The variables included in each model and the relative distribution of risk factors for each dataset are shown in Table 8. Of note, the incidence of age more than 55 years and presence of comorbid disease is relatively stable among the groups, while the distribution of other risk factors varies among the groups. The results of these models is shown in Table 9. When analyzing for "all MOF", five independent risk factors were identified: age more than 55 years emerged in four models (adjusted odds ratios: 2.6, 4.0, 3.7, and 3.7), ISS greater than or equal to 25 in one model (odds ratio: 2.9), 0 to 12 hours RBC transfusion > 6 units in four models (odds ratios: 4.1, 5.6, 3.5, and 4.3), 0 to 12 hours base deficit > 8 mEq/L in two models (odds ratios: 3.1 and 2.6), and 13 to 24 hours lactate > 2.5 mmol/L in one model (odds ratio: 5.1). Using the same stratification, the independent risk factors for "early MOF" included: ISS greater than or equal to 25 emerged in one model (odds ratio: 4.2), ED SBP < 90 mm Hg in one model (odds ratio: 2.6), 0 to 12 hours RBC transfusion > 6 units in four models (odds ratios: 14.5, 37.1, 11.5, and 18.7), and 13 to 24 hours lactate > 2.5 mmol/L one model (odds ratio: 11.7). In contrast, the independent risk factors for "late MOF" included: age more than 55 years in four models (odds ratios: 2.8, 4.8, 4.2, and 4.4), 0 to 12 hours RBC transfusion > 6 units in four models (odds rations: 2.5, 3.1, 2.4, and 2.6), 0 to 12 hours base deficit > 8 mEq/L in one model (odds ratio: 3.2) and 13 to 24 hours lactate > 2.5 mEq/L in one model (odds ratio: 3.6). Of interest, in the "early MOF" models, age did not emerge as an independent risk factor and the odds ratios of the shock indicators were much higher than in the "all MOF" and "late MOF" models. In contrast, in the "late MOF" models, age more than 55 years remains an independent risk factor and the odds ratios of the shock indicators are similar to the "All MOF" models.
Distibution of the number of patients who developed MOF by when they first met the definition of MOF.
Pattern of organ failure upon presentation for early and late MOF.
Distribution of risk factors in various sets of patient data.
Risk factors for MOF, early MOF, and late MOF in sequential patient data sets identified by stepwise logistic regression.
The potential impact of major infections on the clinical course of early MOF compared to late MOF is shown in Table 10. Thirty-two major infections occurred in 23 of the 27 early MOF patients (85%). By our classification, two (6%) were not related, two (6%) were potential "triggers," five (16%) worsened MOF, and 23 (72%) were symptoms. In contrast, 59 major infections occurred in 38 of the 43 late MOF patients (88%). Nineteen (32%) were not related, 16 (27%) appeared to "trigger" late MOF, three (5%) significantly worsened late MOF, and the remaining 21 (36%) appeared to be late symptoms. Thus, early and late MOF patients experienced a similar high incidence of major infectious complications. The majority of these occurred either early and, thus were unrelated, or occurred late and appeared to be symptoms. However, in the late MOF patients, major infections were classified as MOF triggers more often than in early MOF (16 (27%) vs. 2 (6%), p = 0.02). In regards to other outcome variables, no differences were found between early MOF and late MOF patients mechanical ventilation days (16.3 +/- 2.4 vs. 20.0 +/- 1.9, p = 0.24), ICU stay (21.8 +/- 3.4 vs. 28.0 +/- 2.8 days, p = 0.17) or overall mortality (12 of 27 patients (44%) vs. 13 of 43 patients (30%), p = 0.34). Survival curves from the onset of MOF for early MOF and late MOF are shown in Figure 5. While the overall curves are not statistically different (Wilcoxon Test, p = 0.23), it appears that early MOF patients die in a shorter time period from the onset of MOF.
Classification of major infection in early and late MOF.
Survivor curves of early versus late MOF.
DISCUSSION
The MOF syndrome emerged two decades ago as a result of our ability to keep critically ill patients alive by advanced ICU technology. [1-3] Early clinical studies identified infection as being a prime inciting event. [2,3,21-23] Based on these observations, two infectious models of MOF were proposed: (1) insult forward arrow ARDS forward arrow pulmonary sepsis forward arrow MOF or (2) insult forwar arrow sepsis (principally intra-abdominal abscess) forward arrow ARDS/MOF. With these models in mind, researchers focused their attention on determining how a traumatic event sets the stage for infection and how infections mechanistically drive MOF. [16-19] As a result, throughout the 1980s, prevention, diagnosis and treatment of postinjury infections improved, but secondary infections remained a common MOF associated event. [36,37] The epidemiology of these infections appeared to change and in large part their continued high rate was owing to failure of local and systemic host defenses. Pneumonia remained clinically important, while intra-abdominal abscess became a less common event. Simultaneously, European reports of blunt trauma patients demonstrated that MOF could occur in the absence of infection. [24,25] Subsequently, alternative hypotheses were proposed suggesting trauma could induce a malignant systemic inflammatory response syndrome (SIRS) that precipitates MOF independent of infection (e.g., bacterial translocation, ischemia/reperfusion and "one-hit" and "two-hit" inflammatory models). [16-19] At the basic level, these proposed mechanisms are well documented and compelling. Unfortunately, we have not been able to translate this knowledge into the clinical practice. In part, we believe this is because of inadequate epidemiologic characterization of the MOF syndrome. [20]
We have focused our studies on trauma patients for several reasons. First, early postinjury SIRS appears to play a pivotal role in the ultimate pathogenesis of MOF and the mechanisms involved in early postinjury SIRS appear to be different from those involved in infection-driven SIRS. [30,38] Secondly, the inciting traumatic event can be identified and characterized independent of the resulting SIRS. Additionally, the early SIRS response can in turn be characterized independent of the outcome MOF. In other patient populations traditionally associated with MOF (e.g., gram-negative sepsis and intra-abdominal infection) these events are not temporally distinct. The inciting event is frequently diagnosed by the SIRS response and severe SIRS at diagnosis may overlap with the outcome variable MOF. [39-41] Additionally, trauma patients tend to be young and free of preexisting disease, while in other patient populations, advanced age and comorbidity are significant confounders. [31,32] Lastly, as a result of regionalized trauma care, major trauma patients are triaged to trauma centers where care is delivered by standardized protocols. This is invaluable not only when studying epidemiology, but also in the design of focused observational studies and ultimately interventional trials.
Interestingly, in a recent comprehensive study of trauma death in San Diego County, it was concluded that optimization of this regionalized trauma system lowered the incidence of sepsis and multiple organ failure. [6] Thirty-one (14%) of 224 in-hospital were owing to pneumonia or sepsis, which is lower than the 43 (21%) of 205 in-hospital deaths resulting from sepsis reported by Baker et al. in 1980. [4] Thus, if this comparison is valid, the incidence of MOF-related deaths is decreasing. However, this should not be construed to mean that the incidence of MOF is decreasing. In fact, the incidence of MOF appears to be increasing and the confusion arises because fewer MOF patients are dying. Given the different definitions and denominators used, this is a difficult issue to prove. However, in six recent studies (including ours), the incidence of MOF in high-risk trauma patients ranges from 14% to 42%, which is higher than the 7% reported by Fry et al. in 1980 and the 8% reported by Faist et al. in 1983. [11-15,22,24] This is an important trauma system issue. Severely injured patients who do not die early because of optimal initial management are admitted to ICUs where they receive expensive, high-acuity care. In the current study, 15% of trauma ICU patients who survived more than 48 hours, developed MOF. This cohort of patients required a mean of 19 days of mechanical ventilation and remained in our ICU for a mean of 26 days. Thus, from a resource perspective, MOF remains a significant problem.
The primary observation of the current study is that postinjury MOF appears to occur in at least two different patterns (i.e., early versus late). This concept is not new. In fact, in a 1975 report from Denver, Walker and Eiseman [2] noted that the pattern of postinjury ARDS was changing. Of 78 trauma patients requiring mechanical ventilation, 13 (17%) developed ARDS. Of these, nine had classic early onset ARDS (within 12 hours) and all survived. The remaining five patients developed late ARDS (more than 5 days), all were septic and all died of pulmonary insufficiency or bacteremic hypotension. The presentation was so disparate, the authors concluded that they were dealing with different diseases. In 1983, Faist et al., [24] in an often-quoted Bavarian study, similarly noted two patterns of MOF. Of the 433 blunt polytrauma patients studied, 34 (8%) developed MOF. In 15 (44%), the onset was rapid (12 to 36 hours), apparently the result of combined severe multisystem trauma and shock. In the remaining 19 (56%), the onset was late (average 7.2 days) and uniformly associated with sepsis. Again in 1992, Waydhas et al. [13] from Munich, in a prospective study of 100 severe multisystem-injured patients (mean ISS = 37), noted that 45 developed organ failure within 2 days (primarily ARDS), and 14 evolved into MOF. A second peak of late MOF (predominated by liver failure) emerged in another 18 patients at 6 to 8 days. In nine (50%) of these late MOF patients, infections immediately preceded or coincided with onset of MOF. These findings are quite similar to our recent study of postinjury ventilator-associated pneumonias. [35] In this prospective study of 123 high-risk torso trauma patients (mean ISS = 36 +/- 2) who required more than 24 hours of mechanical ventilation, 28 patients (23%) developed MOF. In 14 (50%), the onset was early (less than or equal to3 days). Eleven of these patients developed pneumonias; in four cases, the onset was temporally associated with worsening MOF, while the remaining seven cases occurred late and had no significant impact on MOF scoring (i.e., they appeared to be "symptoms"). In the other 14 patients, the onset of MOF was late (more than 3 days). Nine of these patients developed pneumonia, and in eight, [57] the diagnosis of pneumonia was temporally associated with the onset of MOF (i.e., they appeared to be "triggers").
Collectively, the above studies corroborate our current study and are consistent with the hypothesis that MOF occurs as a result of a dysfunctional inflammatory response (see Figure 6). [19,41-47] Following major trauma, patients are resuscitated into an early state of hyperinflammation (i.e., SIRS). This can result in early MOF in the initial insult is massive ("one-hit" model) or if early secondary inflammatory insults occur ("two-hit" model). Alternatively, as time proceeds, negative feedback systems down-regulate early SIRS to limit potential autodestructive inflammation. This results in delayed immunosuppression, which is associated with major infectious complications. If this hypothesis is accurate, one strategy would be to limit early hyperinflammation. Our basic and clinical research studies suggest that the postinjury SIRS becomes activated as early as 6 to 12 hours postinjury. [48-54] Thus, anti-inflammatory interventions may need to be initiated quite early. Additionally, postinjury SIRS is complex and involves multiple effector cells with overlapping mediator cascades. Mild to moderate SIRS is most likely beneficial (i.e., it is the normal "injury stress response"), while severe SIRS is potentially harmful. Unfortunately, our current knowledge is limited in identifying which components should be modulated to achieve a favorable outcome. The recent problems with various anti-inflammatory strategies in large clinical studies for gram-negative sepsis, supports the need for observational studies in trauma. [48-50]
Global hypothesis for the pathogenesis of postinjury MOF.
An alternative approach to modulating early hyperinflammation is to focus on delayed immunosuppression and the associated late infections. One crucial question is whether late infections cause ongoing MOF and thus contribute to mortality. The epidemiologic studies of Border et al., [52] Marshall et al., [53] and other provide compelling evidence to invoke the gut as the potential reservoir for pathogens in late MOF. Additionally, several gut-specific preventative strategies (e.g., early enteral feeding, immune-enhancing diets and selective gut decontamination) have been shown to reduce postinjury infections. [54-58] While the nutrition studies lack sufficient patient numbers to assess mortality as an outcome variable, there have been sufficiently large selective gut decontamination (SGD) trials to address this issue. Multiple studies have shown that SGD reduces major infections (principally pneumonia), but it does not consistently reduce mortality. [56,57] This has lead to the hypothesis that late infections are inconsequential symptoms of MOF. [58] The trauma-related SGD trials, however, may be flawed because they enrolled a large portion of head trauma patients whose mortality is not linked to sepsis-related MOF. [59] Therefore, we have two options with these existing studies: either we accept reduced infections as a reasonable goal or we need to perform epidemiologic studies to identify study candidates whose infections have attributable mortality and then enroll enough of these patients in new interventional trials to assess mortality as an endpoint. Another cogent issue is whether late MOF-associated "pneumonia" truly reflects active infection or are we just treating immuno-compromised hosts who are heavily colonized. Patients with late ARDS generally have signs of infections (e.g., low-grade fever, leukocytosis), chest roentgenographs that are heavily infiltrated and thus noninterpretable, and are likely to have endotracheal tube aspirate cultures positive as a result of prolonged intubation. In the last 4 years, we have been using bronchoalveolar lavage (BAL) to assist in this clinical dilemma. [60] In select patients with refractory ARDS, who have failed to respond to antibiotic therapy, we stop antibiotics and perform a BAL. In those patients whose clinical course does not deteriorate and the quantitative culture of the BAL fluid does not yield heavy growth, we treat with high-dose corticosteroids. We have observed dramatic improvement in over 80% of these patients. [61] Others observed similar success. [62,63] Perhaps more late MOF patients might benefit by this seemingly paradoxical therapy.
In conclusion, this study has documented that MOF remains a significant problem in our regionalized trauma system. Additionally, MOF appears to present in at least two patterns (i.e., early versus late). Better understanding of the relative roles of dysfunctional inflammation and infections in early MOF versus late MOF may facilitate the development of new strategies for the prevention and treatment of this morbid syndrome.
Acknowledgments
The authors express their sincere gratitude to Diane Blackmon for her excellent preparation of the manuscript.
PAPER DISCUSSION
Dr. Donald E. Fry (Albuquerque, New Mexico): I really appreciate the opportunity to discuss another fine presentation by Dr. Moore and his associates from Denver.
The issue of multiple organ failure is obviously one of considerable interest to everyone here that takes care of critically injured trauma patients. Unfortunately, the literature on this subject over the last 15 to 18 years is littered with different definitions, different criteria, dramatically different outcomes, very different observations by a whole host of very capable and qualified people.
It thus makes comparisons between these numerous reports virtually impossible. Everyone seems to use a different definition of what multiple organ failure is. The patient populations that are studied are different. The organs that are studied are indeed different.
So if I would take the liberty of comparing our own work of some 1,200 patients studied about 12 to 15 years ago with that of Dr. Moore's today, I find out that using our criteria compared to his criteria, that he's identifying twice as many multiple organ failure patients with half the mortality rate that we did about 15 years ago.
So one could reach a conclusion, if you just took things at face value, that multiple organ failure's not declining; it's increasing in frequency, since they saw it in 15% of the patients admitted in this series, but at only half the mortality rate.
Well, obviously, I don't believe that there's twice as many organ failure patients today, nor do I believe that there's half the mortality rate.
One of the things I would raise in this particular presentation that I'd like Dr. Moore to respond to is the use of heart failure as a criterion for organ dysfunction in these patients. The criteria that's been used is the utilization of inotropes. Inotropes are used for a host of different indications in the clinical arena, many times in patients that are having a severe systemic inflammatory response. The inotropes actually end up being used to deal with systemic vascular resistance problems and have absolutely nothing to do whatsoever with dysfunction or poorly functioning myocardium.
So I would really like Dr. Moore to assuage my concerns that he was dealing with peripheral vascular problems and that the heart was actually functioning very well, and accordingly he ended up with more organ failure patients than in fact really existed.
I've noticed that he's totally eliminated the issue of stress bleeding as a potential multiple organ failure expression, and I would be curious if Dr. Moore could address for us whether stress bleeding has in fact totally disappeared from the high altitude of Denver, and whether prophylaxis against that measure is indeed the cause for that?
I was interested to see the early and late multiple organ failure expressions, since I would have considered his late group actually to be early, and would have expected there to be an ultralate group that would have fell out some 14, 18, 21 days later, and that was clearly not seen in his group.
The patients that were identified in the early MOF group looked to me to be the patients that have had inadequate resuscitation, not because of care problems, but simply because of profound injury, inability to resuscitate. It would seem to me that those early 72-hour MOF patients were in fact patients that had severe shock and ischemia, and I'm not really surprised that there were a significant number of patients identified there.
On the other hand, the late MOF patients identified in this study only really had 27% of the patients have infection identified as a triggering event. I'm curious as to what was the frequency of the systemic inflammatory response syndrome in this group of patients. I think most of us, even those of us who advocated infection as the critical trigger in organ failure, no appreciate that it really is, I believe, the activation of the systemic inflammatory response syndrome, which may or may not be triggered by infection.
So I think a more meaningful observation. Fred, might be of those late organ failure patients, what number of those really did have SIRS.
Finally, I guess I continue to be troubled by why are we doing these kinds of studies, and I'd like for Dr. Moore to address that issue. When we all continue to use different criteria, we cannot correlate our information with each other. When we change the definitions, we can't longitudinally evaluate whether we're making improvements or not in the care.
I guess my final plea in this discussion is that we must come to some kind of a standard nomenclature in multiple organ failure if we're ever going to be able to make any sense out of this morass of data and publications. Thank you very much.
Dr. Philip Bosco (Granite Bay, California): I would just ask Dr. Moore what do you call a patient--a trauma patient--who has had bad shock, and has had the obligatory one or two blood volume transfusions, has ARDS, and 2 or 3 days later his or her bilirubin is eight or ten from blood transfusion? Is that multiple organ failure, sequential organ failure, or is that just a byproduct of blood transfusion?
Dr. Ronald J. Simon (Bronx, New York): I enjoyed your presentation very much. I just have one question. There's a rising interest in the lungs as a motor organ for multiorgan failure, and I just wonder how many of your patients with late organ failure had ARDS prior to their onset of multiorgan failure? Thank you.
Dr. H. Gill Cryer (Los Angeles, California): I'd like to amplify on the last question a little bit. Since the score is cumulative, it seems that what you may really have is two different rates of progression of the disease, both starting on the day of admission, rather than a group that develops it late and one that develops it earlier. In other words, the late group may actually start accumulating MOF points on day 1 but not accumulate enough points to meet the MOF threshold until day 7.
So, like the last questioner, is there evidence of a single organ dysfunction early on that progresses more slowly in the late MOF group?
If that is the case, did infection then lead to a progression toward a more severe syndrome?
Dr. Philip S. Barie (New York, New York): A very nice presentation, Fred. I enjoyed it. I have two questions, if I may.
First, you use a reasonable, if arbitrary, definition of multiple organ failure (MOF) and say that your patients either have it or they do not. We have been interested in quantitating MOF on a continuous basis rather than as an all or nothing phenomenon. If you change your cutoff point or look at a continuum of scores, does that influence your data?
Second, in your early patients who manifest MOF can you, at some point, discriminate outcomes relatively early on? Thank you.
Dr. Frederick A. Moore (closing): Dr. Fry, we aggressively resuscitate trauma patients at known risk for MOF. By 24 hours, virtually all the patients who are not going to die early (i.e., within 48 hours of admission) are hyperdynamic. At this juncture we curtail resuscitative efforts and persist only in patients who have evidence of a persistent or recurrent peripheral perfusion deficit. We do not intentionally maintain supernormal oxygen delivery beyond acute resuscitation nor do we attempt normalize systemic vascular resistance (SVR). We believe that a decreased SVR is the normal compensatory response in a severely injured patient who has been adequately resuscitated. We intentionally do not obtain organ dysfunctions scores until day 3 to distance ourselves from acute resuscitation efforts and at this time define heart failure as a cardiac index less than 3 on moderate dose of inotropes. Secondly, significant stress gastritis bleeding has disappeared. I cannot remember the last time a trauma patient at DGH needed operative intervention for this entity. In my experience, most high risk patients who are endoscoped early will have grade I to III gastric erosions. However, despite different prophylactic regimens, these do not progress to clinical bleeding. Compared to your experience in the late 1970s, the natural history of this entity has definitely changed. I believe this is because of better overall ICU care. We currently use sucralfate for prophylaxis because it has the least potential side effects. Along the same line, I believe the lack of a third peak of ultralate MOF is because of the changing epidemiology of postinjury intra-abdominal abscesses. Compared to your experience in the late 1970s, far fewer of our patients developed intra-abdominal abscesses. Additionally, they appeared to be diagnosed and treated earlier as a result of the availability of CAT scanning. Moreover their impact on triggering or worsening MOF was less important. Next, while I acknowledge that SIRS is a valuable concept, my problem with SIRS is that its definition is too sensitive. By the standard definition virtually all of our study patients had SIRS. A cogent issue is how to differentiate mild, moderate, and severe SIRS. Our medical colleagues, lead by Dr. Bone, in a recent consensus conference, quantitate the severity of SIRS by the presence or absence of organ dysfunctions. Thus by these criteria, all of our MOF patients have severe SIRS. Finally, the term MOF was coined by Dr. Ben Eiseman at Denver General Hospital in 1977. Since that time, MOF has been a focus of our surgery department's research efforts. Our scoring system was developed 10 years ago. It originally included eight organ dysfunctions; however, four were subsequently dropped because their definitions were subjective and they did not substantially contribute to identifying MOF. I believe that our revised score is as good as any currently available. I agree that a consensus conference, with surgical input, is badly needed.
Dr. Bosco, the scenario you describe is a bit atypical. Two to 3 days after a massive transfusion, the rise in bilirubin secondary to blood products would typically be in the range of only 2 to 3 mg/dL, although occasionally bilirubin will acutely rise above 5 if there was underlying liver disease or a transfusion reaction. On the other hand, liver failure in the MOF syndrome tends to present as a second peak in bilirubin beginning at 5 to 7 days postinjury.
Dr. Simon and Dr. Cryer, by our definition, patients develop MOF when their simultaneous organ dysfunction score exceeds a threshold. In both early and late MOF, the lung is frequently the first organ to fail. Presumably, the lungs are more vulnerable to systemic inflammation or our clinical tools to detect lung dysfunction are more sensitive. Nevertheless, patients rarely die of isolated lung failure. In late MOF, the lungs may indeed serve as a source of ongoing systemic inflammation. This is certainly true for a subset of ARDS patients who develop pneumonia.
Dr. Barie, we agree that defining MOF as a continuous variable from a statistical standpoint is better than defining it as a dicotomized variable. However, this is quite difficult to do since there is no gold standard for MOF. Additionally, the relative contribution of the various organ dysfunctions need to be quantitated. This will require a much larger data base. In regards to your second question, I assume the MOF score at early presentation would be predictive of death since 44% of these patients ultimately died. However, when we developed our MOF score we were not interested in predicting death, rather we were trying to describe the disease, which in turn could serve as an endpoint in our clinical studies.
REFERENCES
1. Baue AE: Multiple progressive, or sequential systems failure. Arch Surg 110:779, 1975
2. Walker L, Eiseman B: The changing pattern of post-traumatic respiratory distress syndrome. Ann Surg 181:693, 1976
3. Eiseman B, Beart R, Norton L: Multiple organ failure. Surg Gynecol Obstet 144:323, 1977
4. Baker CC, Oppenheimer L, Stephens B, et al: Epidemiology of trauma deaths: Am J Surg 140:144, 1980
5. Moore FA, Moore EE: Trauma Resuscitation. In Wilmore DW, Brennan MF, Harken AH, et al (eds): Scientific American Surgery. New York, Scientific American, 1995, pp 1-63
6. Shackford SR, MacKersie RC, Holbrook TL, et al: The epidemiology of traumatic death. Arch Surg 128:571, 1993
7. Sauaia A, Moore FA, Moore EE, et al: Early predictors of postinjury multiple organ failure. Arch Surg 129:39, 1994
8. Sauaia A, Moore FA, Moore EE, et al: Epidemiology of trauma deaths: A reassessment. J Trauma 38:185, 1995
9. Lauwers LF, Roelant A, Beeckman C, et al: A retrospective study of 130 consecutive multiple patients in an intensive care unit. Intensive Care Med 12:296, 1986
10. Crump JM, Duncan DA, Wears R: Analysis of multiple organ system failure in trauma and nontrauma patients. Am J Surg 54:702, 1988
11. Goins WA, Reynolds HN, Nyanjom D et al: Outcome following prolonged intensive care unit stay in multiple trauma patients. Crit Care Med 19:339, 1991
12. Henao FJ, Daes JE, Dennis RJ: Risk factors for multiorgan failure: A case control study. J Trauma 31:75, 1991
13. Waydhas C, Nost-Kolb D, Jochum M, et al: Inflammatory mediators, infection, sepsis, and multiple organ failure after severe trauma. Arch Surg 127:460, 1992
14. Tran DD, Cuesta MA, Leeuwen PA, et al: Risk factors for multiple organ system failure and death in critically injured patients. Surgery 114:21, 1993
15. Regel G, Lobenhoffer P, Grotz M, et al: Treatment results of patients with multiple trauma: An analysis of 3406 cases treated between 1972 and 1991 at a German level I trauma center. J Trauma 38:70, 1995
16. Baue AE (ed): Multiple Organ Failure: Patient Care and Prevention. St. Louis, Mosby Year Book, 1990
17. Fry DE (ed): Multiple System Organ Failure. St. Louis, Mosby Year Book, 1992
18. Demling R, LaLonde C, Saldinger P, et al: Multiple-organ dysfunction in the surgical patient: Pathophysiology, prevention and treatment. Curr Probl Surg 30:347, 1993
19. Moore FA, Moore EE: Evolving concepts in the pathogenesis of postinjury multiple organ failure. Surg Clin North Am 75:257, 1995
20. Sauaia A, Moore FA, Moore EE, et al: Early risk factors for postinjury multiple organ failure. World J Surg. In press
21. Polk HC, Shields CL: Remote organ failure: A valid sign of occult intra-abdominal infection. Surgery 81:310, 1977
22. Fry DE, Pearlstein L, Fulton RL, et al: Multiple system organ failure: The role of uncontrolled infection. Arch Surg 115:136, 1980
23. Bell RC, Coalson JJ, Smith JD, et al: Multiple organ system failure and infection in adult respiratory distress syndrome. Ann Intern Med 99:293, 1983
24. Faist E, Baue AE, Mittmer H, et al: Multiple organ failure in polytrauma patients. J Trauma 23:775, 1983
25. Goris JA, Boekhoerst TP, Nuytinck JK, et al: Multiple organ failure. Arch Surg 120:1109, 1985
26. Montgomery Ab, Stager MA, Carrico J, et al: Cause of mortality in patients with the adult respiratory distress syndrome. Am Rev Respir Dis 132:485, 1985
27. Villar J, Manzano JJ, Blazsquez MA, et al: Multiple system organ failure in acute respiratory failure. J Crit Care 6:75, 1991
28. Murray JF, Mallhay MA, Luce JM, et al: An expanded definition of the adult respiratory distress syndrome. Am Rev Respiratory Dis 138:720, 1988
29. Bernard GR, Artigas A, Brigham KL, et al: Report of the American-European Consensus Conference on Acute Respiratory Distress Syndrome: Definitions, mechanisms, relevant outcome and clinical trial coordination. J Crit Care 8:1, 1993
30. Moore FA, Moore EE, Poggetti R, et al: Gut bacterial translocation via the portal vein: A clinical perspective with major torso trauma. J Trauma 31:629, 1991
31. Knaus WA, Wagner DP, Draper EA, et al: The APACHE III prognostic system: Risk prediction of hospital mortality for critically ill hospitalized adults. Chest 100:1619, 1991
32. Milzman DP, Hinson D, Magnant CM: Trauma and pre-existing diseases: Overview and outcome. Crit Care Clin 9:633, 1993
33. The Abbreviated Injury Scale 1990 Revision. De Plaines, Ill, Association for the Advancement of Automotive Medicine, 1991
34. Gainec JC, Jarvis WR, Horan TC, et al: CDC definitions for nosocomial infections. Am J Infect Control 16:128, 1988
35. Sauaia A, Moore FA, Moore EE, et al: Pneumonia: Cause or symptom of postinjury multiple organ failure? Am J Surg 166:607, 1993
36. Fry DE (ed): A symposium: The continued challenge of trauma and infections in surgery. Am J Surg 165:1S, 1993
37. Stillwell M, and Caplan ES: The septic multiple-trauma patient. Infect Dis Clin North Am 3:155, 1989
38. Hoch RC, Rodriguez R, Manning T, et al: Effects of accidental trauma on cytokine and endotoxin production. Crit Care Med 21:839, 1993
39. Bone RC, Balk RAA, Cerra FB, et al: American College of Chest Physicians/Society of Critical Care Medicine Consensus Conference: Definitions for sepsis and organ failure and guidelines for the use of innovative therapies in sepsis. Crit Care Med 20:864, 1992
40. Bone RC, Balk R, Slotman G, et al: Adult respiratory distress syndrome: Importance of development of multiple organ failure. Chest 10:3320, 1992
41. Moore FA, Peterson VM, and Moore EE: Inadequate granulopoieses after major torso trauma: A hematopoietic regulatory paradox. Surgery 108:667, 1990
42. Moore EE, Moore FA, Francoise RJ, et al: Postischemic gut serves as a priming bed for circulating neutrophils that provoke multiple organ failure. J Trauma 37:881, 1994
43. Botha AJ, Moore FA, Moore EE, et al: Early postinjury neutrophil priming and activation: A vulnerable window. Surgery 118:358, 1995
44. Botha AJ, Moore FA, Moore EE, et al: Early neutrophil sequestration after injury: A pathologic mechanism for multiple organ failure. J Trauma 39:411, 1995
45. Polk HC Jr, George CD, Wellhausen SR, et al: A systematic study of host defense processes in badly injured patients. Ann Surg 204:282, 1986
46. Faist E, Mewes A, Baker CC, et al: Prostaglandin E sup 2 (PGE sup 2)-dependent suppression of interleukin alpha (IL2) production in patients with major trauma. J Trauma 27:837, 1987
47. Chaudry IH, Ayala A: Immunologic Aspects of Hemorrhage. Austin, Tex, RG Landes, 1992
48. Cronin L, Cook DJ, Carlet J, et al: Corticosteroid treatment of sepsis: A critical appraisal and meta-analysis of the literature. Crit Care Med 23:1430, 1995
49. Fisher CJ, Slatman GJ, Opal SM, et al: Initial evaluation of human recombinant interleukin-1 receptor antagonist in the treatment of sepsis syndrome: A randomized, open-label, placebo-controlled multicenter trial. Crit Care Med 22:12, 1994
50. Fisher CJ, Opal SM, Dhainaut J, et al: Influence of an antitumor necrosis factor monoclonal antibody on cytokine levels in patients with sepsis. Crit Care Med 21:318, 1993
51. Bone RC, Bolk RA, Fein AM, et al: A second large controlled clinical study of E%, a monoclonal antibody to endotoxin: Results of a prospective, multicenter, randomized, controlled trial. Crit Care Med 23:994, 1995
52. Border JR, Hassett J, LaDuca J, et al: The gut origin septic states in blunt multiple trauma (ISS = 40) in the ICU. Ann Surg 206:427, 1987
53. Marshall, JC, Christou NV, Meakins JL: The gastrointestinal tract: the "undrained abscess" of multiple organ failure. Ann Surg 218:111, 1993
54. Moore FA, Feliciano DV, Andrassy RJ, et al: Early enteral feeding, compared with parenteral, reduces postoperative septic complications. Ann Surg 216:172, 1992
55. Moore FA, Moore EE, Kudsk KA, et al: Clinical benefits of an immune-enhancing diet for early postinjury enteral feeding. J Trauma 37:607, 1994
56. Van Saene HKF, Stoutenbeek CC, et al: Selective decontamination of the digestive trace in the intensive care unit: Current status and future prospects. Crit Care Med 20:691, 1992
57. Heyland DK, Cook DJ, Jaeschke A, et al: Selective decontamination of the digestive tract: An overview. Chest 105:1221, 1994
58. Cerra FB, Maddaus MA, Dunn DL, et al: Selective gut decontamination reduces nosocomial infections and length of stay but not mortality or organ failure in surgical intensive care unit patients. Arch Surg 127:163, 1992
59. Sauaia A, Moore FA, Moore EE, et al: Pneumonia related multiple organ failure is not a primary cause of death in head trauma. Pan Am J Trauma 3:90, 1992
60. Sauaia A, Moore FA, Moore EE, et al: Diagnosing pneumonia in mechanically ventilated trauma patients: Endotracheal aspirate versus bronchoalveolar lavage. J Trauma 35:512, 1993
61. Biffl WL, Moore FA, Moore EE, et al: Are corticosteroids salvage therapy for refractory ARDS? Am J Surg. 170:591, 1995
62. Ashbaugh DG, Maier RV: Diopathic pulmonary fibrosis in adult respiratory distress syndrome. Arch Surg 120:530, 1985
63. Meduri GU, Chinn AJ, Leeper KV, et al: Corticosteroids rescue treatment of fibroproliferative ARDS. Chest 105:1516, 1994
© Williams & Wilkins 1996. All Rights Reserved.