The systemic inflammatory response syndrome

10 downloads 0 Views 123KB Size Report
Systemic inflammatory response syndrome (SIRS) is the clinical expression of ... SIRS and MODS are graded expressions of the inflammation associated with ...
Journal of Antimicrobial Chemotherapy (1998) 41, Suppl. A, 1–7

JAC

The systemic inflammatory response syndrome: definitions and aetiology Per-Olof Nyström* Department of Surgery, University Hospital, SE-581 85 Linköping, Sweden Systemic inflammatory response syndrome (SIRS) is the clinical expression of the action of complex intrinsic mediators of the acute phase reaction. SIRS can be precipitated by events such as infection, trauma, pancreatitis, and surgery. At times, SIRS can compromise the function of various organ systems resulting in Multiple Organ Dysfunction Syndrome (MODS). SIRS and MODS are graded expressions of the inflammation associated with acute illness. Mild forms are frequent in general wards for both medical and surgical patients but those with severe forms require intensive care. Clinicians should learn to identify SIRS in their patients at an early stage to determine the underlying cause and treatment before the SIRS progresses to a more severe form.

sepsis are shown in Table I. Traditionally, sepsis was taken to mean microbial infection even in the absence of infection proven by culture. The idea of a microbial cause of all cases of ‘sepsis’ is well established, so much so that clinicians may speak of occult infection when the syndrome occurs but bacteria cannot be cultured. Nearly all patients with signs of sepsis are empirically treated with antimicrobial agents, since delaying treatment until definite microbiological evidence of infection is available can result in a higher rate of morbidity/mortality. When such treatment is unsuccessful it is common practice to discontinue antibiotics for a period in the hope that further culture will reveal any microorganisms present. The clinical signs of sepsis, such as fever, tachycardia, tachypnoea and leucocytosis, are common responses to systemic infection. A trigger–response concept of sepsis emerged, in which bacteria were seen as the trigger, with the pathophysiology being the response to that trigger. It was now recognized that sepsis involved both microbial and pathophysiological events, and this powerful concept allowed clinicians to see the role of infection in the common inflammatory response. For example, uninfected trauma patients and those with intra-abdominal infection had similar clinical courses, both groups developing multiple organ failure with identical microscopic pathology—autodestructive inflammation which seemed to be independent of infection.5 Others found that the clinical response persisted after eradication of the infection and was itself associated with increased mortality.6 In animal

Introduction The introduction of the term ‘systemic inflammatory response syndrome’ (SIRS) by the American College of Chest Physicians and Society of Critical Care Medicine (ACCP/SCCM) consensus conference1 recognized the important role that endogenous mediators of systemic inflammation play in ‘sepsis’, which was no longer regarded as being caused by microbial pathogenicity factors alone. This very different concept initiated a re-evaluation in approach to clinical features of the acute severe illness associated with infection.2–4 The similar definitions of sepsis syndrome and SIRS indicate a change of perspective rather than a new clinical entity. In this review sepsis/SIRS will be viewed as a ‘trigger–mediator–response’ sequence, where SIRS is the response to the action of intrinsic mediators. The identification of SIRS does not confirm a diagnosis of infection or sepsis since the features of SIRS can be seen in many other conditions such as trauma, pancreatitis, burns or infection, or following major elective surgery (Figure 1). The pathophysiological responses can be quantified and, if appropriately scored, will provide a probability of mortality.

Previous definitions of sepsis The varying contributions of local pathology, physiological responses and microorganisms made the precise definition of sepsis very difficult. Excerpts from various definitions of

*Tel: +46-13-223550 or -223541; Fax +46-13-223570; E-mail: [email protected]

1 © 1998 The British Society for Antimicrobial Chemotherapy

P.-O. Nyström possibility of systemic response in the absence of invasive infection. Such patients are common among both medical and surgical patients.3 Do these patients have a genuine infection that has failed to be cultured, or a focus of infection that is inaccessible for culture? Either is unlikely since few sources of invasive infection escape detection with modern imaging methods or remain undiagnosed by blood culture. The same decision matrix also includes invasive infection without accompanying systemic inflammation; this clinical situation is sometimes encountered in immunocompromised patients. Thus the trigger may be present without an adequate SIRS response, and a SIRS response can occur in the absence of infection.

Figure 1. The ACCP/SCCM concept of systemic inflammation (SIRS) as a common response to many initiating circumstances. Depending on the severity of the insult patients may or may not display SIRS. Modified from Bone et al.1

The consensus conference definitions of SIRS and MODS

studies it was found that the severity of the physiological response was a better predictor of outcome than the microbial challenge.7,8 These and other studies showed the value of breaking down sepsis into its components which could then be studied separately. The criteria defining sepsis are satisfied when proven microbial infection and a systemic response are both present (Table II). The decision matrix also shows the

The criteria proposed by the ACCP/SCCM are given in Table III; they were derived from those previously used to identify the sepsis syndrome (Table IV). At least two criteria are needed for the identification of SIRS. The consensus conference also recognized a progression in the disease state from simple SIRS/sepsis to severe SIRS/sepsis in the presence of acute organ dysfunction, hypotension or hypoperfusion. The next step in the progression is ‘SIRS

Table I. Some previous definitions of sepsis which demonstrate the development of the concept of sepsis as a combination of infection and pathophysiology Definition of sepsis

Reference

Bacterial infection The presence of at least one positive blood culture concomitant with the presence of a septic focus growing the same microorganism Infection is a process, sepsis is the response The systemic response to infection Infection as a microbiological phenomenon was differentiated from sepsis as a clincial syndrome Sepsis can be defined as composed of two elements; the abnormal presence of microbes, and the host responses by endogenous mediators

Goris et al.5 (J. L. Meakins, personal communication) ACCP/SCCM consensus conference1 Marshall & Sweeney6 Jönsson et al.8

Table II. Decision matrix for the two components of sepsis: invasive infection and systemic inflammation Infection (invasive)

SIRS (systemic inflammation)

Clinical interpretation

Absent Present

present absent

Present

present

culture-negative SIRS, or non-infectious cause infection without sepsis, or failed response, or immunocompromised state sepsis

2

Systemic inflammatory response syndrome Table III. Definitions given by the ACCP/SCCM consensus conference. Modified from Bone et al.1 Term

Definition

SIRS

the systemic inflammatory response to a variety of severe clinical insults. The response is manifested by two or more of the following conditions: temperature 38°C or 36°C; heart rate 90 beats/min; respiratory rate 20 breaths/min or PaCO2 32 torr ( 4.3 kPa); WBC 12,000 cells/mm3, 4000 cells/mm3 or 10% immature (band) forms microbial phenomenon characterized by an inflammatory response to the presence of microorganisms or the invasion of normally sterile host tissue by those organisms the systemic response to infection; this response is manifested by two or more of the SIRS criteria as a result of infection the presence of viable bacteria in the blood sepsis associated with organ dysfunction, hypoperfusion or hypotension; hypoperfusion and perfusion abnormalities may include, but are not limited to, lactic acidosis, oliguria or an acute alteration in mental status sepsis with hypotension, despite adequate resuscitation with fluids, along with the presence of perfusion abnormalities that may include, but are not limited to, lactic acidosis, oliguria or an acute alteration in mental status; patients who are on inotropic or vasopressor agents may not be hypotensive when perfusion abnormalities are measured a systolic blood pressure of 90 mmHg or a reduction of 40 mmHg from baseline in the absence of other causes for hypotension presence of altered organ function in an acutely ill patient such that homeostasis cannot be maintained without intervention

Infection Sepsis Bacteraemia Severe sepsis

Septic shock

Hypotension MODS

Abbreviations: SIRS, systemic inflammatory response syndrome; MODS, multiple organ dysfunction syndrome.

Table IV. The definition of sepsis syndrome after Bone et al.9 Clinical evidence of infection Rectal temperature 38°C or 36°C Tachycardia ( 90 beats/min) Tachypnoea ( 20 breaths/min while spontaneously breathing) At least one of the following manifestations of inadequate organ function/perfusion: alteration in mental status hypoxaemia (PaO2) 72 torr breathing room air) (overt pulmonary disease not direct cause of hypoxaemia) oliguria (urine output 30 mL or 0.5 mL/kg for at least 1 h)

shock’ present in patients with SIRS and hypotension that is unresponsive to resuscitation with fluids. Multiple organ dysfunction syndrome (MODS) is the recognized diminished organ function associated with acute illness, in which organ function is not capable of maintaining homeostasis. The dysfunction may be absolute or relative but is more readily identified as a gradual change over time. An example of relative organ dysfunction is the patient who has normal cardiac output and systemic oxygen delivery but inadequate tissue oxygenation (e.g. lactic acidosis). It is further recognized that organ dysfunction may be a direct result of a well-defined insult. The

secondary form of multiple organ dysfunction is an integral part of the SIRS itself. The consensus conference gave no indication of the organ involvement or degree of dysfunction that identify MODS.

Pathophysiology The definition of sepsis syndrome as developed by Bone et al.9 (Table IV) was originally seen as a diagnostic tool to be used in many clinical studies of sepsis and trials of anti3

P.-O. Nyström infective agents to identify patients thought to have invasive infection.9 It was gradually realized that it was not a diagnostic tool but merely identifies patients who match the definition criteria. It does not identify all patients thought to have sepsis, nor is there a specific prognosis associated with sepsis syndrome. The APACHE III database10 contains information for scoring patients either by the sepsis syndrome or by the definition of the APACHE III outcome model. In a subgroup of 519 ICU patients from the database with sepsis but without specific infection only about 60% had the sepsis syndrome or septic shock. The prognoses for these patients ranged from near zero risk to very high risk, with similar risk distributions for those who satisfied the criteria and those who did not.11 However, the criteria for defining the sepsis syndrome are closer to those of the severe SIRS/sepsis category of the consensus conference: when the basic SIRS criteria were applied nearly all of the 519 patients qualified as having SIRS.1 The SIRS criteria cannot perform much better for diagnosis or as a measure of prognosis, perhaps because they are too wide. A recent prospective survey identified 542–857 episodes of SIRS/1000 patient-days in cardiovascular, medical and surgical intensive care units,3 and in general surgical, cardiothoracic and medical oncology wards 320–671 episodes of SIRS/1000 patient-days were identified. Only about half of the patients had proven infection, the others having culture-negative SIRS. SIRS is common and may alert the clinician to patients who have systemic inflammation which may be explained by the working diagnosis or to those who need further diagnostic or therapeutic measures. SIRS is not a diagnosis nor is it a good indicator of outcome but its presence in individual patients must be explained adequately. It will need to be replaced by a more complete and comprehensive definition, or by terms that describe the pathophysiology of individual cases.

The mediators of the trigger–response concept After an injection of bacteria or endotoxin, certain cytokines appear briefly in circulating blood.12 The classical sequence is tumour necrosis factor (TNF- ) followed by interleukins 1 (IL-1), 6 (IL-6) and 8 (IL-8). These cytokines have been named the pro-inflammatory or alarm cytokines because they appear first. They, and other factors, mediate various responses including the activation of numerous cell populations and release of secondary cytokines and growth factors. It is important to understand whether the same mediator is activated by different triggers and if different mediators are associated with different pathophysiological responses. Does the pathophysiological response to infection differ from the response to other triggers? There are certain differences: infection induces more TNFthan does physical trauma, which releases more IL-6 and IL-8.12–15 Thus primary infection is associated with higher fever than in trauma, probably because of the different balance of mediators produced. Hypovolaemic and hypotensive patients, however, often initially suffer from hypothermia and leucopenia whether triggered by infection or trauma. After resuscitation all such patients develop fever and leucocytosis. Initial differences in mediator patterns result in different clinical presentations, but within hours or days it is clinically no longer possible to associate these responses with specific triggers. Approaching the problem from a different standpoint, attempts to correlate measurements of circulating cytokines with pathophysiological changes and with prognosis have not been entirely successful. The concentrations of these mediators vary widely probably because they have short half-lives in the circulation and most are localized within the inflamed body compartment where they cannot readily be measured. Patients admitted to intensive care with a diagnosis of acute sepsis, i.e. SIRS following presumed infection, often do not have positive blood cultures providing unequivocal proof of invasive infection, nor do they have endotoxaemia. Many do not have detectable TNF- or IL-1 in response to presumed invasive infection. Even elevated IL6 or IL-8 concentrations in circulating blood are not found consistently. However, detectable bacteria, endotoxin or inflammatory mediators are associated with increased mortality.16

The acute phase response This may be seen as the primary part of the systemic inflammatory response. Tissue injury or bacteria at the site will activate complement and induce tissue macrophages, monocytes and other reactive cell elements such as mast cells, endothelial cells and platelets to produce various mediators.17 TNF- and IL-1 are secreted in large amounts and appear in the circulation within 1 h. They have both

Figure 2. Clinical entities defined in the ACCP/SCCM consensus paper showing the progression to more severe forms of sepsis and the development of multiple organ dysfunction.

4

Systemic inflammatory response syndrome Table V. The multiple organ dysfunction score developed by Marshall et al.28

System

Parameter

Respiratory Renal Hepatic Cardiovascular Haematological Neurological

p(O2)/FI(O2) ratio serum creatinine ( mol/L) serum bilirubin ( mol/L) PARa platelet count Glasgow coma score

a

Pressure-adjusted heart rate (PAR) is calculated as heart rate

0

1

300 100 20 10.0 120 15

226–300 101–200 21–60 10.1–15.0 81–120 13–14

Score 2 151–225 201–350 61–120 15.1–20.0 51–80 10–12

3 76–150 351–500 121–240 20.1–30.0 21–50 7–9

4 75 500 240 30.0 20 6

(right atrial (i.e. central venous) pressure/mean arterial pressure).

local and distant effects. Stromal cells at the site of injury release a secondary wave of mediators that augment the inflammatory signal. IL-6 now appears in the circulation and also has systemic effects. The principal alarm cytokines cause fever and stimulate the pituitary to release stress hormones. They also stimulate the liver to synthesize a number of acute phase proteins such as C-reactive protein, fibrinogen and major antiproteases.18 IL-6 is the principal signal for the hepatic acute phase response. At the site of injury numerous mediators cause pain, vasodilation and increased vascular permeability, and attract granulocytes to the site. The signs of local inflammation (pain, erythema, increased temperature and oedema) are easily recognized. In classical experiments Miles et al.19 showed that the early underlying mechanisms were active over a very short period (about 4 h) after which the local responses ceased. The pathophysiological features of inflammation, however, typically take 24–48 h to develop fully. The acute phase response involves additional mechanisms which reverse the inflammatory response. Interleukins 4 and 10 can turn off monocyte/macrophage production of TNF- , IL-1, IL-6 and IL-8 and therefore have potent anti-inflammatory action. In addition, the acute phase response produces antagonists to the TNF receptor and IL-1 receptor, which bind the circulating cytokine or block receptors on target cells. The early events also induce the production of cortisone. Through the combined action of these mechanisms and others the acute phase response is attenuated and the homeostasic mechanisms return to normal. It is clinically important to understand that the proinflammatory state of the acute phase response also initiates anti-inflammatory activity involving mediators. If the insult has been effectively controlled and contained by host defences, or by means of surgery, then the acute systemic inflammation will be attenuated and will clinically resolve. This process takes several days and is best seen in patients after major surgery. Fever resolves after 2–3 days, bowel function is restored after 3–4 days and the patient can soon return to full oral intake.

The extent of the pro- and anti-inflammatory events is proportional to the magnitude of the insult. Constitutional factors, both genetic and acquired, may cause particular patients to overreact or respond inadequately. The proand anti-inflammatory mechanisms are often dysregulated for reasons that are poorly understood.

Systemic inflammation and organ failure are on a continuum The SIRS criteria will serve only to sort patients into those who display at least two of the criteria and therefore have SIRS by definition, and those who do not have SIRS. Further categorization is possible by identifying those who satisfy two, three or all four criteria. Two additional variables, namely organ dysfunction and circulatory status, can be used to characterize severe SIRS or SIRS shock. When several organs fail to maintain their homeostatic function, MODS can be said to have complicated the SIRS. The various stages must be seen as phases in a continuum of increasing disturbance of homeostasis (Figure 2). The SIRS criteria may be seen as a crude stratification system for patients with systemic inflammation as shown in a recent prospective study of the epidemiology of SIRS in medical and surgical patients.3 Mortality was 3% in patients without SIRS but doubled in those with SIRS. Mortality was increased to 10% in those with three positive criteria and to 17% in those matching all four criteria. Death rates were similar for patients with culture-negative SIRS and those with culture-positive SIRS. This was true also for culture-negative septic shock, which was associated with 46% mortality. The same authors also found evidence of an increase in severity of SIRS with time. Half of the patients with two criteria of SIRS developed a third criterion by day 7. Of patients with two SIRS criteria 36% developed sepsis (culture-proven infection) by day 14 and in those with all four criteria 45% developed sepsis by day 14. The time taken for sepsis to appear decreased as more SIRS criteria were met. 5

P.-O. Nyström Several scoring systems have been introduced to assess the physiological response, but the variables must be weighted carefully so that increasing score values correlate with worsening prognosis. Examples of such scoring methods are the acute physiology and chronic health evaluation (APACHE),10,20 the simplified acute physiology score (SAPS II)21 and the mortality probability model (MPM II).22 The APACHE system is the most widely used system for assessing acute physiological disturbances. The pathophysiology scored is that which predicts the greater mortality risk in patients with different initiating events or diagnoses. The latest version, APACHE III, provides updated risk assessments for the first 7 days of illness.23 The various scoring systems were recently reviewed for their ability to identify septic patients.24 As most systems have not been validated extensively their relative merits are difficult to assess. The APACHE stands out as the best validated system. Based on the APACHE III methodology two new sepsis-oriented scores were developed from large populations of septic patients.25,26 However, these scores were developed after sophisticated statistical modelling and they cannot be applied without access to computer support. There are probably no simplistic solutions to predictive scoring of SIRS/sepsis. Simple scores that can be used at the bedside are unreliable in individual patients, while more precise scores are so complex that they are unsuitable for routine use. As so many patients with SIRS have a low mortality risk and are not treated in ICUs, such sophisticated scoring is unnecessary. The clinician needs to recognize the presence of systemic inflammation early to diagnose and treat its cause before the SIRS progresses to a more severe form. The consensus conference gave no specific definition of MODS. Mild degrees of organ dysfunction are common in SIRS but the untreated patient may develop severe organ failure, which is associated with poor outcome. The simultaneous failure of two organs is associated with widely differing prognosis depending on the combination of organs involved, ranging from 20% mortality for combined cardiovascular and haematological failure to 76% mortality for combined cardiovascular and CNS failure.27 A score based simply on the number of failed organ systems is therefore an unreliable predictor, especially in the absence of agreed definitions. The various methods of grading organ dysfunction have been recently reviewed for the development of a new score of organ failure.28 This MOD score (Table V) considers six organ systems. The function of each organ system is weighted on a scale from 0 to 4. The weighting was developed in one half of a surgical ICU population and validated in the other half. For each variable a zero value was associated with 5% mortality while a value of 4 represented severe dysfunction and a mortality of 50% for patients managed in the ICU. This score showed excellent correlation both for patients treated in the ICU and for those in general wards. Since the MOD score is based

on pathophysiology and not on therapeutic components, it should be objective and produce consistent results in different patient populations.

Conclusions Systemic inflammation may be thought of as the presence of signs that are clinically associated with inflammation such as fever, tachycardia or leucocytosis. Other signs such as the appearance of C-reactive protein are better described as the acute phase response. Furthermore, many patients with SIRS display varying degrees of organ dysfunction while some progress to develop multiple organ failure. The clinical signs seen in SIRS arise through many mechanisms, including inflammation, acute phase response, metabolic dysfunction, shock, hypoxia and organ system failure. While some physiological signs can be recognized as primary and others as secondary, they all contribute to disease severity, course and prognosis. The definition of SIRS, while imperfect, implies variable levels of risk depending on the degree of pathophysiological disturbance. The production of anti-inflammatory mediators is triggered by inflammation and these will arrest the SIRS and restore normal homeostasis, assuming that medical intervention such as surgery, intensive care treatment or antibiotics has controlled the initial cause.

References 1. Bone, R. C., Balk, R. A., Cerra, F. B., Dellinger, R. P., Fein, A. M., Knaus, W. A. et al. (1992). Definitions for sepsis and organ failure and guidelines for the use of innovative therapies in sepsis. ACCP/SCCM consensus conference committee. Chest 101, 1644–55. 2. Beal, A. L. & Cerra, F. B. (1994). Multiple organ failure syndrome in the 1990s. Systemic inflammatory response and organ dysfunction. Journal of the American Medical Association 271, 226–33. 3. Rangel-Frausto, M. S., Pittet, D., Costigan, M., Hwang, T., Davis, C. S. & Wenzel, R. P. (1995). The natural history of the systemic inflammatory response syndrome (SIRS): a prospective study. Journal of the American Medical Association 273, 117–23. 4. Pittet, D., Rangel-Frausto, S., Li, N., Tarara, D., Costigan, M., Rempe, L. et al. (1995). Systemic inflammatory response syndrome, sepsis, severe sepsis and septic shock: incidence, morbidities and outcomes in surgical ICU patients Intensive Care Medicine 21, 302–9. 5. Goris, R. J., te Boekhorst, T. P., Nuytinck, J. K. & Gimbrère, J. S. (1985). Multiple-organ failure: generalized autodestructive inflammation? Archives of Surgery 120, 1109–15. 6. Marshall, J. & Sweeney, D. (1990). Microbial infection and the septic response in critical surgical illness. Sepsis, not infection, determines outcome. Archives of Surgery 125, 17–23. 7. Nieuwenhuijzen, G. A., Haskel, Y., Lu, Q., Berg, R. D., van Rooijen, N., Goris, R. J. et al. (1993). Macrophage elimination increases bacterial translocation and gut-origin septicemia but attenuates symptoms and mortality rate in a model of systemic inflammation. Annals of Surgery 218, 791–9.

6

Systemic inflammatory response syndrome 8. Jönsson, B., Berglund, J., Skau, T. & Nyström, P. O. (1993). Outcome of intra-abdominal infection in pigs depends more on host responses than on microbiology. European Journal of Surgery 159, 571–8.

19. Miles, A. A., Miles, E. M. & Burke, J. (1957). The value and duration of defence reactions of the skin to the primary lodgement of bacteria. British Journal of Experimental Pathology 38, 79–96. 20. Knaus, W. A., Draper, E. A., Wagner, D. P. & Zimmerman, J. E. (1985). APACHE II: a severity of disease classification. Critical Care Medicine 13, 818–29.

9. Bone, R. C., Fisher, C. J., Clemmer, T. P., Slotman, G. J., Metz, C. A. & Balk, R. A. (1989). The sepsis syndrome: a valid clinical entity. Critical Care Medicine 17, 389–93.

21. Le Gall, J. R., Lemeshow, S. & Saulnier, F. (1993). A new Simplified Acute Physiology Score (SAPS II) based on a European/North American multicenter study. Journal of the American Medical Association 270, 2957–63.

10. Knaus, W. A., Wagner, D. P., Draper, E. A., Zimmerman, J. E., Bergner, M., Bastos, P. G. et al. (1991). The APACHE III prognostic system. Risk prediction of hospital mortality for critically ill hospitalized adults. Chest 100, 1619–36.

22. Lemeshow, S., Teres, D., Klar, J., Avrunin, J. S., Gehlbach, S. H., Rapoport, J. et al. (1993). Mortality Probability Models (MPM II) based on an international cohort of intensive care unit patients. Journal of the American Medical Association 270, 2478–86.

11. Knaus, W. A., Sun, X., Nyström, P. O. & Wagner, D. P. (1992). Evaluation of definitions for sepsis. Chest 101, 1656–62. 12. Michie, H. R., Spriggs, D. R., Manogue, K. R., Sherman, M. L., Revhaug, A., O’Dwyer, S. T. et al. (1988). Tumor necrosis factor and endotoxin induce similar metabolic responses in human beings. Surgery 104, 280–6.

23. Wagner, D. P., Knaus, W. A., Harrell, F., Zimmerman, J. E. & Watts, C. (1994). Daily prognostic estimates for critically ill adults in intensive care units. Critical Care Medicine 22, 1359–72.

13. Bone, R. C. (1996). Toward a theory regarding the pathogene sis of the systemic inflammatory response syndrome: what we do and do not know about cytokine regulation. Critical Care Medicine 24, 163–72.

24. Barriere, S. L. & Lowry, S. F. (1995). An overview of mortality risk prediction in sepsis. Critical Care Medicine 23, 376–93.

14. Dinarello, C. A. & Cannon, J. G. (1993). Cytokine measurements in septic shock. Annals of Internal Medicine 119, 853–4.

25. Knaus, W. A., Harrell, F. E., Fisher, C., Wagner, D. P., Opal, S. M. Sadoff, J. C. et al. (1993). The clinical evaluation of new drugs for sepsis: a prospective study design based on survival analysis. Journal of the American Medical Association 270, 1233–41.

15. Roumen, R. M., Hendriks, T., van der Ven-Jongekrijg, J., Nieuwenhuijzen, G. A., Sauerwein, R. W., van der Meer, J. et al. (1993). Cytokine patterns in patients after major vascular surgery, hemorrhagic shock, and severe blunt trauma. Relation with subsequent adult respiratory distress syndrome and multiple organ failure. Annals of Surgery 218, 769–76.

26. Knaus, W. A., Harrell, F. E., LaBreque, J. F., Wagner, D. P., Pribble, J. P., Draper, A. E. et al. (1996). Use of predicted risk of mortality to evaluate the efficacy of anticytokine therapy in sepsis. Critical Care Medicine 24, 46–56.

16. Casey, L. C., Balk, R. A. & Bone, R. C. (1993). Plasma cytokines and endotoxin levels correlate with survival in patients with sepsis syndrome. Annals of Internal Medicine 119, 771–8.

27. Zimmerman, J. E., Knaus, W. A., Wagner, D. P., Sun, X., Hakim, R. B. & Nyström, P. O. (1996). A comparison of risks and outcomes for patients with organ system failure: 1982–1990. Critical Care Medicine 24, 1633–41.

17. Baumann, H. & Gauldie, J. (1994). The acute phase response. Immunology Today 15, 74–80.

28. Marshall, J. C., Cook, D. J., Christou, N. V., Bernard, G. R., Sprung, C. L. & Sibbald, W. J. (1995). The multiple organ dysfunction score: a reliable descriptor of a complex clinical outcome. Critical Care Medicine 23, 1638–52.

18. Steel, D. M. & Whitehead, A. S. (1994). The major acute phase reactants: C-reactive protein, serum amyloid P component and serum amyloid A protein. Immunology Today 15, 81–8.

7