© 2006 American Thoracic Society DOI: 10.1165/rcmb.F308
Pathobiology of SepsisAre We Still Asking the Same Questions?Division of Pulmonary and Critical Care Medicine, Division of Infectious Diseases and Channing Laboratory; and Department of Medicine, Brigham and Women's Hospital, Boston Massachusetts Correspondence and requests for reprints should be addressed to Rebecca M. Baron, M.D., Division of Pulmonary and Critical Care Medicine, Brigham and Women's Hospital, 75 Francis Street, Boston, MA 02115. E-mail: rbaron{at}partners.org
While the term "sepsis" has been used for centuries to describe critical illness resulting from infection, a specific definition and conceptual basis for its genesis are still debated (1–3). Despite incredible clinical technological advances and molecular progress in defining the pathobiology of bacterial infection, the incidence of sepsis and associated mortality remain exceedingly high (4). To date, with the exception of activated protein C, no recently developed specific therapies for sepsis have proven effective, despite numerous clinical trials (5). Failures of these trials, many of which targeted identified mediators of the inflammatory cascade, led to cries for reinvestigation of basic disease mechanisms (3, 6). In fact, recent studies have elicited exciting new answers to very old questions, which suggest novel avenues of investigation and future treatment strategies for sepsis. In this review, we set out to explore how our understanding of the pathobiology of sepsis has evolved, through examining historically how investigators have tried to answer four basic mechanistic questions:
Sepsis as "Rottenness": Theory of Spontaneous Generation The concept of "sepsis" dates at least as far back as the time of Hippocrates and was viewed as a process of dangerous, odorous biologic decay, or putrefaction (7). Ancient Greeks believed this process was one form of biologic breakdown that could occur within the body and, furthermore, that there existed "dangerous principles" within the colon that were capable of causing "autointoxication." Aristotle, and later the Romans, proposed that within swamps, the process of sepsis resulted in the production of invisible living creatures that emanated in putrid fumes termed miasmata (Figures 1 and 2). Thus persisted for centuries the concept of "spontaneous generation" of tiny dangerous animals, and early public health initiatives in Roman cities were directed toward elimination of these perceived deadly swamps.
Only in 1684 did a controlled experiment begin to refute the concept of "spontaneous generation." Francisco Redi baited three pots with meat, which was allowed to "age" under the following conditions (8):
In the first condition, flies flew in and out of the pot freely, and the meat was flooded with maggots shortly thereafter. In the second condition, maggots were seen only upon the gauze, not in contact with the meat. In the third condition, the meat in the airtight container was free of maggots. Redi concluded that rather than generating the insects, rotten meat attracted the flies to lay the eggs that produced the maggots. While this experiment was compelling, the contemporaneous development of the microscope confirmed the presence of "simple" tiny animals, leading to the revival of "spontaneous generation" as an explanatory model for sepsis—a model that persisted for nearly 200 years more.
Sepsis as a Result of Infection: Development of "The Germ Theory of Disease" Suddenly a thought crossed my mind: childbed fever and the death of Professor Kolletschka were one and the same. His sepsis and childbed fever must originate from the same source...the fingers and hands of students and doctors, soiled by recent dissections, carry those death-dealing cadavers' poisons into the genital organs of women in childbirth.... Semmelweiss instituted handwashing precautions before patient contact, and this intervention resulted in a reduction in the overall mortality rate from 18% to less than 3% in his maternity ward. However, his results were not well received by the medical community, as reflected in this statement from an American obstetrician: "Doctors are gentlemen, and gentlemen's hands are clean." Semmelweiss was fired from his position. He later died in an insane asylum—ironically, from an illness resembling puerperal sepsis, incurred from a finger laceration (9). An elegant experiment in 1859 by Louis Pasteur finally disproved "spontaneous generation." He allowed boiled broth to stand in flasks that had curved swan necks, such that microbes from the air could settle on the necks of the flask, but broth could contact the microbes only if the flask was tipped to bring broth into the neck of the flask. In this famous experiment, he found that sterile broth not in contact with microbes exhibited no bacterial growth when allowed to sit for prolonged periods, while sterile broth that contacted microbes on the neck of the flask always turned cloudy. Thus, Pasteur conclusively demonstrated that the process of putrefaction required living organisms (8). Furthermore, in the 1870s, Robert Koch, a country physician, showed that a particular microbe was the etiologic agent of anthrax, a reproducible clinical syndrome that was afflicting large numbers of farmers and livestock in his community. Koch later described the etiologic agents of a number of other diseases, including cholera and tuberculosis, and received a Nobel prize in recognition of these achievements 100 years ago (10). His work, in conjunction with Pasteur's findings, established the "germ theory of disease," finally supplanting "spontaneous generation." Koch's postulates for microbial disease causation established the field of microbiology (8):
On the basis of these principles, Joseph Lister contemporaneously demonstrated the beneficial effects of antiseptic techniques (use of carbolic acid wound treatments), in significantly reducing perioperative mortality from limb amputation (11). His work, along with the writings of Pasteur and Koch, ultimately ushered in the conceptual notion of infection as causal in sepsis (7). However, the scientific community was still slow to adapt, as he writes: ...during the previous nine months, in which the antiseptic system had been fairly in operation in my wards, not a single case of pyaemia, erysipelas, or hospital gangrene had occurred in them... but the carrying out of this rule implies a conviction of the truth of the germ theory of putrefaction, which, unfortunately, is in this country the subject of doubts such as I confess surprise me, considering the character of the evidence which has been adduced in support of it.... The concept of sepsis as the systemic response to infection has persisted, and remains defined as such to this day based upon the deliberations of multidisciplinary consensus conferences (1, 2) (Table 1). However, ongoing debate and discussion persist in defining sepsis, as reflected in proposed revisions to the 1992 consensus conference definitions, which will be discussed below (2). Interestingly, the definitions of sepsis and related conditions currently in use are remarkably similar to those put forth by William Osler in The Principles and Practice of Medicine 100 years earlier (1, 12) (Table 1). Furthermore, Osler's proclamation of the critical nature of source control in sepsis management remains true to this day (12, 13): "The brilliant and remarkable results which follow complete evacuation of pus with thorough drainage give the indication of the only successful treatment of this condition."
Visualizing Bacteria While is it now appreciated that infection and sepsis can arise from nonbacterial pathogens, we have focused this historical review on bacterial pathogens. As described above, even before the development of the microscope and the ability to visualize bacteria, ancient people imagined that there existed invisible creatures, or "small animals (animalcules)" that were responsible for dangerous diseases. Anton von Leeuwenhoek first described bacterial organisms (cocci, rods, and spirals) in 1674 using a single-lens microscope, but a systematic study of bacteria was not undertaken until nearly 100 years later. In fact, a drawing of "small animals (animalcules)" made by Louis Joblot in 1718 suggests the appearance of protozoa (including one with a human face) (Figure 1A) (7, 14). Thus resolution of commercially available lenses was likely not sufficient for uniform visualization of bacteria at that time.
Culturing Bacteria
The acceptance of "the germ theory," the successful application of antiseptic techniques in the nineteenth century, and treatment of infections with antibiotics in the twentieth century all supported the concept that bacteria caused infections that were detrimental to the host. However, several observations suggest that there are still gaps in our understanding of the pathobiology of sepsis. First, while early administration of appropriate antibiotics (within 1 h of presentation) improves sepsis outcomes and remains a crucial part of the current standard of care (16, 17), a significant mortality rate persists in sepsis, despite treatments targeting the microbe(s) identified in cultures (4). A number of explanations have been proposed for this phenomenon, including delay in appropriate antibiotic or resuscitative treatment (17), increased virulence of the existing microbes (18, 19), or exuberant uncontrollable host responses (which will be discussed below). Second, cultures remain negative in
Molecular Identification of Bacteria Recently, examination of rDNA sequences was undertaken to identify vaginal bacterial species from patients with bacterial vaginosis (BV), a disease in which causative organisms have not been identifiable in vaginal fluid cultures (23). Instead, diagnosis has been made traditionally on clinical grounds or with Gram's staining, and the condition has been poorly responsive to antibiotic treatment. Interestingly, the investigators found greater diversity of bacterial species identified by 16S rDNA analysis of vaginal fluid from patients with BV than from control patients. Furthermore, patients with BV were found to carry 16 novel bacterial species not previously described, which were not isolated from the control group patients. Further, fluorescence in situ hybridization demonstrated that these new bacterial organisms resembled those often visualized by Gram's staining in vaginal fluid of patients with BV. This association study does not prove that these novel bacterial species cause BV, and BV clearly represents a different clinical scenario from sepsis (24). However, two small studies have explored use of rDNA analysis in identifying causative bacterial pathogens in neonatal sepsis (25). These studies raise the tantalizing possibility of rapid, increasingly specific pathogen identification techniques in the future that may dramatically improve our ability to identify pathogens against which to target antibacterial therapy.
The concept of a "toxic" factor in sepsis dates back to ancient times, as described above, and has recurred in different conceptual phases throughout history. Herodotus reported that ancient Egyptians used enemas three times per month to purge themselves of the dangerous toxic substances in the gastrointestinal system (4). In more recent times, physicians caring for wounded soldiers during World War I recognized a subset of patients with hypotension disproportionate to the extent of the blood loss. Furthermore, these patients were refractory to the standard resuscitation measures used at the time (26). Animal models of hind-limb injury studied by Walter Cannon and others in 1918–1920 suggested that a toxic substance (histamine) was absorbed from the injured muscle that propagated a shock-like state independent of the extent of blood loss. However, improved animal models and more advanced techniques applied by Alfred Blalock disproved this concept in 1930, and he was able to attribute the entirety of the shock-like state to extensive blood loss within the injured limb. Regardless, debate persisted regarding the existence of a specific "toxic" factor propagating the same common pathway of demise in all forms of human shock. The advent of antibiotic use in the 1940s resulted in a shift from gram-positive to gram-negative organisms as the predominant bacterial pathogen isolates, which persisted for the ensuing 40–50 years (27, 28). Although the concepts of bacterial toxins and endotoxemia had existed since the late 1800s, it was predominantly in the 1950s–1960s that increasing frequencies of infections with gram-negative rods produced a clinical picture deemed consistent with endotoxemic shock (29). The similar clinical presentation of patients with refractory traumatic shock and with endotoxemic shock reinvigorated inquiry toward whether a final common pathway mediated death in these two populations of critically ill patients. Interestingly, speculation again focused on the gastrointestinal system as a source of demise, with use of animal models to explore splanchnic hypoperfusion as an instrumental process in terminal shock from any cause (30). However, in the late 1960s, it became increasingly appreciated that patients with coliform bacterial infections suffered from a clinical process distinct from that of refractory traumatic shock. Francis Moore's writings in 1969 (29) foreshadow the development of the 1992 consensus conference definitions of the Systemic Inflammatory Response Syndrome (SIRS: the systemic activation of the innate immune system, regardless of cause) and sepsis (SIRS + infection) (Table 1): Unfortunately for our campaign to eliminate the word "shock" from the surgical lexicon and thus help to untangle the confusion between sepsis and trauma there is no other monosyllable that quite does the job. Still, we would be better off without the word, and our teaching would be clearer if we never used it, since such a departure would cause each surgeon or physician to specify causes and mechanisms rather than contenting himself with the diagnosis of "shock," implying that all animals and patients in such a state are essentially the same. Until the late 1980s, sepsis was considered by many to be near-synonymous with gram-negative endotoxemia (Figure 2) (27, 31). In the early 1990s, a number of studies led to the concept that sepsis can arise from microbial infections in the absence of endotoxin:
These studies formed the groundwork for the breadth of molecular studies over the last 10 years that argue against a single critical toxic factor in sepsis. Rather, this work points to a number of bacterial virulence factors, or "microbial-associated molecular patterns," that can trigger the host response through specific "pattern recognition receptors." These topics have been reviewed in detail recently (4, 33) and are summarized in Figure 2.
Until relatively recently, the preponderance of pathobiological mechanisms proposed to explain sepsis has focused upon the course that the pathogen has taken in a ravaged bystander-host. Ancient peoples believed that epidemics were inflicted upon powerless humans as punishment for wrongdoing (34). In 1892, Osler wrote that the outcome from sepsis was dependent upon the "dose" of the poison from the bacteria and the likelihood of whether a focus of infection could be evacuated (12). Even so, Osler struggled with how to intervene upon the host response: Fever alone is not, I think, hurtful, but it is difficult to differentiate the effects of fever and of the poisons circulating in the blood. It is not impossible, as some suppose, that the fever be directly beneficial; still, high and prolonged pyrexia is undoubtedly dangerous and should be combated.
This debate over the beneficial versus harmful effect of the host response to infection has persisted until today. The notion of sepsis as an uncontrolled proinflammatory response gained support with the identification of proinflammatory cytokines, most notably tumor necrosis factor (TNF)- In the last 10 years, remarkable progress has been made in characterizing the critical role of the host and its interaction with the pathogen in the pathobiology of sepsis. There has been major advancement in our understanding of molecular signaling pathways in response to microbial pathogens, including identification of Toll-like receptors, Nod 1 and Nod 2 intracellular pattern recognition receptors, and the peptidoglycan recognition proteins. Additional information is emerging regarding the systemic effects of the inflammatory cell response on numerous target genes that influence cellular and subcellular processes, including the microcirculation and mitochondrial function. These and other recent developments have been reviewed in detail (4, 17, 35) and are outlined in Figure 2.
The past century has produced a remarkable explosion of knowledge in the pathobiology of sepsis. Compared with over 2,000 years of belief in the "spontaneous generation" theory, in the last 100 years we have progressed from the adoption of the "germ theory" of disease to delineating the molecular pathways of the host-pathogen response. Despite these discoveries, our current definition of sepsis remains remarkably similar to that from the time of Osler (Table 1), and little progress has been made in effective, targeted therapy. To make treatment advances, we will need to reformulate our notion of sepsis from that of a general concept (i.e., the host response to infection), to one of specific diseases characterized by individual pathogens and recognizable host-response profiles (3, 6). Toward this goal, the 2003 consensus conference proposed a staging system for sepsis (modeled after the TNM system for cancer), called the PIRO system, as a "template for future investigation" (Table 1) (1). This system would stratify patients based on their Predisposition (e.g., immunosuppressed state, genetic polymorphisms predicting response to sepsis [4]), Insult infection (e.g., specific pathogens isolated by culture or bacterial DNA, presence of endotoxin), Response (e.g., presence of shock, levels of inflammatory markers, levels of protein C or TNF-  that might predict response to specific types of treatment), and Organ dysfunction (e.g., organ failure composite score, measures of cellular response to insult such as apoptosis or mitochondrial dysfunction). Thus, while we are still asking the same questions regarding the pathobiology of sepsis as we did 100 years ago, our answers are changing and are yielding exciting new discoveries. It is our hope that the next centennial review of sepsis in the year 2105 will recount the successful application of molecular knowledge toward the development of effective treatment strategies for sepsis.
The authors are grateful to Drs. Andrew Onderdonk and Dan Milner (Brigham and Women's Hospital) for providing the bacterial culture plate and Gram's stain images, respectively (Figure 1B). The authors also thank Drs. Lior Dolgonos, Scott Schissel, and Laura Fredenburgh for helpful comments.
Conflict of Interest Statement: None of the authors has a financial relationship with a commercial entity that has an interest in the subject of this manuscript. Accepted in final form December 13, 2005
This article has been cited by other articles:
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Top
Introduction
"What Is Sepsis?"
30% of patients with sepsis. While negative cultures are often attributed to prior use of antibiotics (
