INTRODUCTION AND DEFINITIONS
Sepsis is a common and deadly disease. More than two millennia ago, Hippocrates wrote that sepsis was characterized by rotting flesh and festering wounds. Several centuries later, Galen described sepsis as a laudable event required for wound healing. Once the germ theory was proposed by Semmelweis, Pasteur, and others in the nineteenth century, sepsis was recast as a systemic infection referred to as “blood poisoning” and was thought to be due to pathogen invasion and spread in the bloodstream of the host. However, germ theory did not fully explain sepsis: many septic patients died despite successful removal of the inciting pathogen. In 1992, Bone and colleagues proposed that the host, not the germ, was responsible for the pathogenesis of sepsis. Specifically, they defined sepsis as a systemic inflammatory response to infection. Yet sepsis arose in response to many different pathogens, and septicemia was neither a necessary condition nor a helpful term. Thus, these investigators instead proposed the term severe sepsis to describe cases where sepsis was complicated by acute organ dysfunction and the term septic shock for a subset of sepsis cases that were complicated by hypotension despite adequate fluid resuscitation along with perfusion abnormalities.
In the past 20 years, research has revealed that many patients develop acute organ dysfunction in response to infection but without a measurable inflammatory excess (i.e., without the systemic inflammatory response syndrome [SIRS]). In fact, both pro- and anti-inflammatory responses are present along with significant changes in other pathways. To clarify terminology and reflect the current understanding of the pathobiology of sepsis, the Sepsis Definitions Task Force in 2016 proposed the Third International Consensus Definitions specifying that sepsis is a dysregulated host response to infection that leads to acute organ dysfunction. This definition distinguishes sepsis from uncomplicated infection that does not lead to organ dysfunction, a poor course, or death. In light of the wide variation in the ways that septic shock is identified in research, clinical, or surveillance settings, the Third International Consensus Definitions further specified that septic shock be defined as a subset of sepsis cases in which underlying circulatory and cellular/metabolic abnormalities are profound enough to substantially increase mortality risk.
To aid clinicians in identifying sepsis and septic shock at the bedside, new “Sepsis-3” clinical criteria for sepsis include (1) a suspected infection and (2) acute organ dysfunction, defined as an increase by two or more points from baseline (if known) on the sequential (or sepsis-related) organ failure assessment (SOFA) score (Table 297-1). Criteria for septic shock include sepsis plus the need for vasopressor therapy to elevate mean arterial pressure to ≥65 mmHg with a serum lactate concentration >2.0 mmol/L despite adequate fluid resuscitation.
TABLE 297-1Definitions and Criteria for Sepsis and Septic Shock ||Download (.pdf) TABLE 297-1 Definitions and Criteria for Sepsis and Septic Shock
|Condition ||Definition ||Common Clinical Features ||Criteria in 1991/2003 (“Sepsis-1”/”Sepsis-2”) ||Criteria in 2016 (“Sepsis-3”) |
|Sepsis ||A life-threatening organ dysfunction caused by a dysregulated host response to infection ||Include signs of infection, with organ dysfunction, plus altered mentation; tachypnea; hypotension; hepatic, renal, or hematologic dysfunction ||Suspected (or documented) infection plus ≥2 systemic inflammatory response syndrome (SIRS) criteriaa ||Suspected (or documented) infection and an acute increase in ≥2 sepsis-related organ failure assessment (SOFA) pointsb |
|Septic shock ||A subset of sepsis in which underlying circulatory and cellular/metabolic abnormalities lead to substantially increased mortality risk ||Signs of infection, plus altered mentation, oliguria, cool peripheries, hyperlactemia ||Suspected (or documented) infection plus persistent arterial hypotension (systolic arterial pressure, <90 mmHg; mean arterial pressure, <60 mmHg; or change in systolic by >40 mmHg from baseline ||Suspected (or documented) infection plus vasopressor therapy needed to maintain mean arterial pressure at ≥65 mmHg and serum lactate >2.0 mmol/L despite adequate fluid resuscitation |
Sepsis can arise from both community-acquired and hospital-acquired infections. Of these infections, pneumonia is the most common source, accounting for about half of cases; next most common are intraabdominal and genitourinary infections. Blood cultures are typically positive in only one-third of cases, while many cases are culture negative at all sites. Staphylococcus aureus and Streptococcus pneumoniae are the most common gram-positive isolates, while Escherichia coli, Klebsiella species, and Pseudomonas aeruginosa are the most common gram-negative isolates. In recent years, gram-positive infections have been reported more often than gram-negative infections, yet a 75-country point-prevalence study of 14,000 patients on intensive care units (ICUs) found that 62% of positive isolates were gram-negative bacteria, 47% were gram-positive bacteria, and 19% were fungi.
The many risk factors for sepsis are related to both the predisposition to develop an infection and, once infection develops, the likelihood of developing acute organ dysfunction. Common risk factors for increased risk of infection include chronic diseases (e.g., HIV infection, chronic obstructive pulmonary disease, cancers) and immunosuppression. Risk factors for progression from infection to organ dysfunction are less well understood but may include underlying health status, preexisting organ function, and timeliness of treatment. Age, sex, and race/ethnicity all influence the incidence of sepsis, which is highest at the extremes of age, higher in males than in females, and higher in blacks than in whites. The differences in risk of sepsis by race are not fully explained by socioeconomic factors or access to care, raising the possibility that other factors, such as genetic differences in susceptibility to infection or in the expression of proteins critical to the host response, may play a role.
The incidences of sepsis and septic shock depend on how acute organ dysfunction and infection are defined as well as on which data sources are studied. Disparate estimates come from administrative data, prospective cohorts with manual case identification, and large electronic health-record databases. Organ dysfunction is often defined by the provision of supportive therapy, in which case epidemiologic studies count the “treated,” rather than the actual, incidence. In the United States, recent cohort studies using administrative data suggest that upwards of 2 million cases of sepsis occur annually. Shock is present in ~30% of cases, resulting in an estimated 230,000 cases in a recent systematic review. An analysis of data (both clinical and administrative) from 300 hospitals in the United Healthcare Consortium estimated that septic shock occurred in 19 per 1000 hospitalized encounters. The incidences of sepsis and septic shock are also reported to be increasing (according to ICD9-CM diagnosis and procedure codes), with a rise of almost 50% in the past decade. However, the stability of objective clinical markers (e.g., provision of organ support, detection of bacteremia) over this period in a two-center validation study suggests that new ICD-9 coding rules, confusion over semantics (e.g., septicemia versus severe sepsis), rising capacity to provide intensive care, and increased case-finding confound the interpretation of serial trends. Studies from other high-income countries report rates of sepsis in the ICU similar to those in the United States.
While the data demonstrate that sepsis is a significant public-health burden in high-income countries, its impact on the populations of low- and middle-income countries is probably even more substantial because of the increased incidence of infectious diseases and the high prevalence of HIV in some parts of the developing world. Although there are fewer high-quality studies on sepsis in these countries, the available data support sepsis as a major public-health problem. For example, a study of one cohort in rural Uganda found an incidence of laboratory-confirmed sepsis tenfold that of current global sepsis estimates; as only a minority of patients with sepsis develop bacteremia, the incidence of sepsis in the cohort was probably even higher. Case–fatality rates in low- and middle-income countries are also higher than those in high-income countries, as exemplified by two observational cohorts in Brazil with mortality rates >40%.
For many years, the clinical features of sepsis were considered the result of an excessive inflammatory host response (SIRS). More recently, it has become apparent that infection triggers a much more complex, variable, and prolonged host response than was previously thought. The specific response of each patient depends on the pathogen (load and virulence) and the host (genetic composition and comorbidity), with different responses at local and systemic levels. The host response evolves over time with the patient’s clinical course. Generally, proinflammatory reactions (directed at eliminating pathogens) are responsible for “collateral” tissue damage in sepsis, whereas anti-inflammatory responses are implicated in the enhanced susceptibility to secondary infections that occurs later in the course. These mechanisms can be characterized as an interplay between two “fitness costs”: direct damage to organs by the pathogen and damage to organs stemming from the host’s immune response. The host’s ability to resist as well as tolerate both direct and immunopathologic damage will determine whether uncomplicated infection becomes sepsis.
Initiation of Inflammation
Over the past decade, our knowledge of pathogen recognition has increased tremendously. Pathogens activate immune cells by an interaction with pattern recognition receptors (Fig. 297-1), of which four main classes are prominent: Toll-like receptors (TLRs), RIG-I-like receptors, C-type lectin receptors, and NOD-like receptors; the activity of the last group occurs partially in protein complexes called inflammasomes. The recognition of structures conserved across microbial species—so-called pathogen-associated molecular patterns (PAMPs)—by all these receptors results in upregulation of inflammatory gene transcription and initiation of innate immunity. A common PAMP is the lipid A moiety of lipopolysaccharide (LPS or endotoxin), which attaches to the LPS-binding protein on the surface of monocytes, macrophages, and neutrophils. LPS is transferred to and signals via TLR4 to produce and release cytokines such as tumor necrosis factor that grow the signal and alert other cells and tissues. Up to 10 TLRs have been identified in humans.
Select mechanisms implicated in the pathogenesis of sepsis-induced organ and cellular dysfunction. The host response to sepsis involves multiple mechanisms that lead to decreased oxygen delivery (DO2) at the tissue level. The duration, extent, and direction of these interactions are modified by the organ under threat, host factors (e.g., age, genetic characteristics, medications), and pathogen factors (e.g., microbial load and virulence). The inflammatory response is typically initiated by an interaction between pathogen-associated molecular patterns (PAMPs) expressed by pathogens and pattern recognition receptors expressed by innate immune cells on the cell surface (Toll-like receptors [TLRs] and C-type lectin receptors [CLRs]), in the endosome (TLRs), or in the cytoplasm (retinoic acid inducible gene 1–like receptors and nucleotide-binding oligomerization domain–like receptors [NLRs]). The resulting tissue damage and necrotic cell death lead to release of damage-associated molecular patterns (DAMPs) such as uric acid, high-mobility group protein B1, S100 proteins, and extracellular RNA, DNA, and histones. These molecules promote the activation of leukocytes, leading to greater endothelial dysfunction, expression of intercellular adhesion molecule (ICAM) and vascular cell adhesion molecule 1 (VCAM-1) on the activated endothelium, coagulation activation, and complement activation. This cascade is compounded by macrovascular changes such as vasodilation and hypotension, which are exacerbated by greater endothelial leak tissue edema, and relative intravascular hypovolemia. Subsequent alterations in cellular bioenergetics lead to greater glycolysis (e.g., lactate production), mitochondrial injury, release of reactive oxygen species, and greater organ dysfunction.
At the same time, these receptors also sense endogenous molecules released from injured cells—so-called damage-associated molecular patterns (DAMPs), such as high-mobility group protein B1, S100 proteins, and extracellular RNA, DNA, and histones. The release of DAMPs during sterile injuries such as those incurred during trauma gives rise to the concept that the pathogenesis of multiple-organ failure may be similar in sepsis and noninfectious critical illness. In addition to activating the proinflammatory cytokines, the inflammatory responses implicated in the pathogenesis of sepsis also activate the complement system, platelet-activating factor, arachidonic acid metabolites, and nitric oxide.
Sepsis is commonly associated with coagulation disorders and frequently leads to disseminated intravascular coagulation. Abnormalities in coagulation are thought to isolate invading microorganisms and/or to prevent the spread of infection and inflammation to other tissues and organs. Excess fibrin deposition is driven by coagulation via tissue factor, a transmembrane glycoprotein expressed by various cell types; by impaired anticoagulant mechanisms, including the protein C system and antithrombin; and by compromised fibrin removal due to depression of the fibrinolytic system. Coagulation (and other) proteases further enhance inflammation via protease-activated receptors. In infections with endothelial predominance (e.g., meningococcemia), these mechanisms can be common and deadly.
Although the mechanisms that underlie organ failure in sepsis are only partially known, impaired tissue oxygenation plays a key role. Several factors contribute to reduced oxygen delivery in sepsis and septic shock, including hypotension, reduced red-cell deformability, and microvascular thrombosis. Inflammation can cause dysfunction of the vascular endothelium, accompanied by cell death and loss of barrier integrity, giving rise to subcutaneous and body-cavity edema. An excessive and uncontrolled release of nitric oxide causes vasomotor collapse, opening of arteriovenous shunts, and pathologic shunting of oxygenated blood from susceptible tissues. In addition, mitochondrial damage due to oxidative stress and other mechanisms impairs cellular oxygen utilization. The slowing of oxidative metabolism, in parallel with impaired oxygen delivery, reduces cellular O2 extraction. Yet energy (i.e., ATP) is still needed to support basal, vital cellular function, which derives from glycolysis and fermentation and thus yields H+ and lactate. With severe or prolonged insult, ATP levels fall beneath a critical threshold, bioenergetic failure ensues, toxic reactive oxygen species are released, and apoptosis leads to irreversible cell death and organ failure. The actual morphologic changes in sepsis-induced organ failure are also complex. Generally, organs such as the lung undergo extensive microscopic changes, while other organs may undergo rather few histologic changes. In fact, some organs (e.g., the kidney) may lack significant structural damage while still having significant tubular-cell changes that impair function.
The immune system harbors humoral, cellular, and neural mechanisms that may exacerbate the potentially harmful effects of the proinflammatory response. Phagocytes can switch to an anti-inflammatory phenotype that promotes tissue repair, while regulatory T cells and myeloid-derived suppressor cells further reduce inflammation. The so-called neuroinflammatory reflex may also contribute: sensory input is relayed through the afferent vagus nerve to the brainstem, from which the efferent vagus nerve activates the splenic nerve in the celiac plexus, with consequent norepinephrine release in the spleen and acetylcholine secretion by a subset of CD4+ T cells. The acetylcholine release targets α7 cholinergic receptors on macrophages, reducing proinflammatory cytokine release. Disruption of this neural-based system by vagotomy renders animals more vulnerable to endotoxin shock, while stimulation of the efferent vagus nerve or α7 cholinergic receptors attenuates systemic inflammation in experimental sepsis.
Patients who survive early sepsis but remain dependent on intensive care occasionally demonstrate evidence of a suppressed immune system. These patients may have ongoing infectious foci despite antimicrobial therapy or may experience the reactivation of latent viruses. Multiple investigations have documented reduced responsiveness of blood leukocytes to pathogens in patients with sepsis; these findings were recently corroborated by post-mortem studies revealing strong functional impairments of splenocytes harvested from ICU patients who died of sepsis. Immune suppression was evident in the lungs as well as the spleen; in both organs, the expression of ligands for T cell–inhibitory receptors on parenchymal cells was increased. Enhanced apoptotic cell death, especially of B cells, CD4+ T cells, and follicular dendritic cells, has been implicated in sepsis-associated immune suppression and death. In a cohort of >1000 ICU admissions for sepsis, secondary infections developed in 14% of patients, and the associated genomic response at the time of infection was consistent with immune suppression, including impaired glycolysis and cellular gluconeogenesis. The most common secondary infections included catheter-related bloodstream infections, ventilator-associated infections, and abdominal infections. What is not yet understood is the optimal way to identify those sepsis patients who have hyperinflamed rather than immunosuppressed phenotypes. Similarly, it is unknown whether the dysfunctional immune system is driving organ dysfunction and secondary infections or whether the immune system itself is just another dysfunctional organ.
APPROACH TO THE PATIENT Sepsis and Septic Shock
At the bedside, a clinician begins by asking, “Is this patient septic?” Consensus criteria for sepsis and septic shock agree on core diagnostic elements, including suspected or documented infection accompanied by acute, life-threatening organ dysfunction. If infection is documented, the clinician must determine the inciting cause and the severity of organ dysfunction, usually by asking: “What just happened?” Severe infection can be evident, but it is often quite difficult to recognize. Many infection-specific biomarkers and molecular diagnostics are under study to help discriminate sterile inflammation from infection, but these tools are not commonly used. The clinician’s acumen is still crucial to the diagnosis of infection. Next, the primary physiologic manifestations of organ dysfunction can be assessed quickly at the bedside with a six-organ framework, yielding the SOFA score. Particular focus should then be placed on the presence or absence of shock, which constitutes a clinical emergency. The general manifestations of shock include arterial hypotension with evidence of tissue hypoperfusion (e.g., oliguria, altered mental status, poor peripheral perfusion, or hyperlactemia).
The specific clinical manifestations of sepsis are quite variable, depending on the initial site of infection, the offending pathogen, the pattern of acute organ dysfunction, the underlying health of the patient, and the delay before initiation of treatment. The signs of both infection and organ dysfunction may be subtle. Guidelines provide a long list of potential warning signs of incipient sepsis (Table 297-1). Once sepsis has been established and the inciting infection is assumed to be under control, the temperature and white blood cell (WBC) count often return to normal. However, organ dysfunction typically persists.
Two of the most commonly affected organ systems in sepsis are the respiratory and cardiovascular systems. Respiratory compromise classically manifests as acute respiratory distress syndrome (ARDS), defined as hypoxemia and bilateral infiltrates of noncardiac origin that arise within 7 days of the suspected infection. ARDS can be classified by Berlin criteria as mild (PaO2/FiO2, 201–300 mmHg), moderate (101–200 mmHg), or severe (≤100 mmHg). A common competing diagnosis is hydrostatic edema secondary to cardiac failure or volume overload. Although traditionally identified by elevated pulmonary capillary wedge measurements from a pulmonary artery catheter (>18 mmHg), cardiac failure can be objectively evaluated on the basis of clinical judgment or focused echocardiography.
Cardiovascular compromise typically presents as hypotension. The cause can be frank hypovolemia, maldistribution of blood flow and intravascular volume due to diffuse capillary leakage, reduced systemic vascular resistance, or depressed myocardial function. After adequate volume expansion, hypotension frequently persists, requiring the use of vasopressors. In early shock, when volume status is reduced, systemic vascular resistance may be quite high with low cardiac output; after volume repletion, however, this picture may rapidly change to low systemic vascular resistance and high cardiac output.
Acute kidney injury (AKI) is documented in >50% of septic patients, increasing the risk of in-hospital death by six- to eightfold. AKI manifests as oliguria, azotemia, and rising serum creatinine levels and frequently requires dialysis. The mechanisms of sepsis-induced AKI are incompletely understood. AKI may occur in up to 25% of patients in the absence of overt hypotension. Current mechanistic work suggests that a combination of diffuse microcirculatory blood-flow abnormalities, inflammation, and cellular bioenergetic responses to injury contribute to sepsis-induced AKI beyond just organ ischemia.
Typical central nervous system dysfunction presents as coma or delirium. Imaging studies typically show no focal lesions, and electroencephalographic findings are usually consistent with nonfocal encephalopathy. Sepsis-associated delirium is considered a diffuse cerebral dysfunction caused by the inflammatory response to infection without evidence of a primary central nervous system infection. Consensus guidelines recommend delirium screening with valid and reliable tools such as the Confusion Assessment Method for the Intensive Care Unit (CAM-ICU) and the Intensive Care Delirium Screening Checklist (ICDSC). Critical-illness polyneuropathy and myopathy are also common, especially in patients with a prolonged course. For survivors of sepsis, neurologic complications can be severe. In a national (U.S.) representative prospective cohort of >1000 elderly patients with severe sepsis, moderate to severe cognitive impairment increased by 10.6 percentage points among patients who survived severe sepsis (odds ratio, 3.34; 95% confidence interval [CI], 1.53–7.25) over that among survivors of nonsepsis hospitalizations. Many of these limitations persisted for up to 8 years.
Many other abnormalities occur in sepsis, including ileus, elevated aminotransferase levels, altered glycemic control, thrombocytopenia and disseminated intravascular coagulation, adrenal dysfunction, and sick euthyroid syndrome. Adrenal dysfunction in sepsis is widely studied and is thought to be related more to reversible dysfunction of the hypothalamic–pituitary axis or tissue glucocorticoid resistance than to direct damage to the adrenal gland. The diagnosis is difficult to establish. Recent clinical practice guidelines do not recommend use of the adrenocorticotropic hormone stimulation test or determination of the plasma cortisol level to detect relative glucocorticoid insufficiency.
Laboratory and Physiologic Findings
A variety of laboratory and physiologic changes are found in patients with suspected infection who are at risk for sepsis. In a 12-hospital cohort of electronic health records related to >70,000 encounters (Fig. 297-2), only tachycardia (heart rate, >90 beats per min) was present in >50% of encounters; the most common accompanying abnormalities were tachypnea (respiratory rate, >20 breaths per min), hypotension (systolic blood pressure, ≤100 mmHg), and hypoxia (SaO2, ≤90%). Leukocytosis (WBC count, >12,000/μL) was present in fewer than one-third of patients and leukopenia (WBC count, <4000/μL) in fewer than 5%. Notably, many features that may identify acute organ dysfunction, such as platelet count, total bilirubin, or serum lactate level, are measured in only a small minority of at-risk encounters. If measured, metabolic acidosis with anion gap may be detected, as respiratory muscle fatigue occurs in sepsis-associated respiratory failure. Other, less common findings include serum hypoalbuminemia, troponin elevation, hypoglycemia, and hypofibrinogenemia.
Distribution of SIRS and SOFA variables among infected patients at risk for sepsis, as documented in the electronic health record. Dark green bars represent the proportion of such patients with abnormal findings; light green bars, the proportion with normal findings; and white bars, the proportion with missing data. (Adapted from CW Seymour et al: Assessment of clinical criteria for sepsis: For the Third International Consensus Definitions for Sepsis and Septic Shock [Sepsis-3]. JAMA 315:762, 2016.)
There is no specific test for sepsis, nor is there a gold-standard method for determining whether a patient is septic. In fact, the definition of sepsis can be written as a logic statement:
sepsis = f (threat to life | organ dysfunction | dysregulated host response | infection),
where sepsis is the dependent variable, which in turn is a function of four independent variables linked in a causal pathway, with—from left to right—one conditional upon the other. There may be uncertainty about whether each variable exists, whether it can be measured, and whether the causal and conditional relationships hold. If we assume that organ dysfunction exists and can be measured, then attributing the marginal degradation in function to a dysregulated host response is not simple and requires the ability to determine preexisting dysfunction, other noninfectious contributions to organ dysfunction, and—ideally—the mechanism by which the host response to an infection causes organ dysfunction.
In order to sort through these complex details, clinicians need simple bedside criteria to operationalize the logic statement (Fig. 297-3). With this mandate, the Sepsis Definitions Task Force recommended that, once infection is suspected, clinicians consider whether it has caused organ dysfunction by determining a SOFA score. The SOFA score ranges from 0 to 24 points, with up to 4 points accrued across six organ systems. The SOFA score is widely studied in the ICU among patients with infection, sepsis, and shock. With ≥2 new SOFA points, the infected patient is considered septic and may be at ≥10% risk of in-hospital death.
Schematic of the importance of accurate, easy-to-use criteria for sepsis and its components, infection and organ dysfunction. In the ideal case (left), criteria clearly distinguish sepsis patients from other patients with uncomplicated infection or organ dysfunction. The reality (right), however, is that existing criteria fail to make clear distinctions, leaving a significant proportion of patients in areas of uncertainty. (Adapted from DC Angus et al: A framework for the development and interpretation of different sepsis definitions and clinical criteria. Crit Care Med 44:e113, 2016.)
Because the SOFA score requires multiple laboratory tests and may be costly to measure repeatedly, the quick SOFA (qSOFA) score was proposed as a clinical prompt to identify patients at high risk of sepsis outside the ICU, whether on the medical ward or in the emergency department. The qSOFA score ranges from 0 to 3 points, with 1 point each for systolic hypotension (≤100 mmHg), tachypnea (≥22 breaths/min), or altered mentation. A qSOFA score of ≥2 points has a predictive value for sepsis similar to that of more complicated measures of organ dysfunction. The qSOFA score is undergoing broader evaluation in other cohorts, in low-and middle-income settings, and in algorithms linked to clinical decision-making. Recent work has also shown that, although SIRS criteria may be fulfilled in sepsis, they sometimes are not and do not meaningfully contribute to the identification of patients with suspected infection who are at greater risk of a poor course, ICU admission, or death—outcomes more common among patients with sepsis than among those without.
As stated above, recent definitions have specified that septic shock is a subset of sepsis in which circulatory and cellular/metabolic abnormalities are profound enough to substantially increase mortality risk, but the application of this definition as a criterion for enrollment of patients varies significantly in clinical trials, observational studies, and quality improvement work. For clarity, criteria are proposed for septic shock that include (1) sepsis plus (2) the need for vasopressor therapy to elevate mean arterial pressure to ≥65 mmHg, with (3) a serum lactate concentration >2.0 mmol/L after adequate fluid resuscitation.
The new definitions and diagnostic criteria were externally validated in >1 million encounters stored in electronic health records. Nevertheless, given the uncertainty around the diagnosis of sepsis, Sepsis-3 is undergoing both validation in prospective studies and incorporation into clinical practice and quality improvement initiatives.
Arterial lactate is a long-studied marker of tissue hypoperfusion, and hyperlactemia and delayed lactate clearance are associated with a greater incidence of organ failure and death in sepsis. In a study of >1200 patients with suspected infection, 262 (24%) of 1081 patients exhibited an elevated lactate concentration (≥2.5 mmol/L) even in the setting of normal systolic blood pressure (>90 mmHg) and were at elevated risk of 28-day in-hospital mortality. However, lactic acidosis may occur in the presence of alcohol intoxication, liver disease, diabetes mellitus, administration of total parenteral nutrition, or antiretroviral treatment, among other conditions. Furthermore, in sepsis, an elevated lactate concentration may simply be the manifestation of impaired clearance. These factors may confound the use of lactate as a stand-alone biomarker for the diagnosis of sepsis; thus it should be used in the context of other markers of infection and organ dysfunction.
TREATMENT Sepsis and Septic Shock Early Treatment of Sepsis and Septic Shock
Recommendations for sepsis care begin with prompt diagnosis. Recognition of septic shock by a clinician constitutes an emergency in which immediate treatment can be life-saving. Up-to-date guidelines for treatment are derived from international clinical practice guidelines provided by the Surviving Sepsis Campaign. This consortium of critical care, infectious disease, and emergency medicine professional societies has issued three iterations of clinical guidelines for the management of patients with sepsis and septic shock (Table 297-2).
The initial management of infection requires several steps: forming a probable diagnosis, obtaining samples for culture, initiating empirical antimicrobial therapy, and achieving source control. More than 30% of patients with severe sepsis require source control, mainly for abdominal, urinary, and soft-tissue infections. The mortality rate is lower among patients with source control than among those without, although the timing of intervention is debated. For empirical antibiotic therapy (Table 297-3), the appropriate choice depends on the suspected site of infection, the location of infection onset (i.e., the community, a nursing home, or a hospital), the patient’s medical history, and local microbial susceptibility patterns. In a single-center study of >2000 patients with bacteremia, the number of patients who needed to receive appropriate antimicrobial therapy in order to prevent one patient death was 4.0 (95% CI, 3.7–4.3).
Antibiotic delays may be deadly. For every 1-h delay among patients with sepsis, a 3–7% increase in the odds of in-hospital death is reported. Although meta-analyses report conflicting results, international clinical practice guidelines recommend the administration of appropriate broad-spectrum antibiotics within 1 h of recognition of severe sepsis or septic shock. Empirical antifungal therapy should be administered only to septic patients at high risk for invasive candidiasis.
The treatment elements listed above form the basis for two “bundles” of care: an initial management bundle to be completed within 3 h of presentation and a management bundle to be completed within 6 h. The initial management bundle includes (1) early administration of appropriate broad-spectrum antibiotics, (2) collection of blood for culture before antibiotic administration, and (3) measurement of serum lactate levels. The management bundle includes (1) an intravenous fluid bolus, (2) treatment with vasopressors for persistent hypotension or shock, and (3) re-measurement of serum lactate levels. Implementation of these two bundles has been associated with improved outcome in large multinational studies.
Other elements of the initial management bundle are cardiorespiratory resuscitation and mitigation of the immediate threats of uncontrolled infection. Early resuscitation requires a structured approach including the administration of IV fluids and vasopressors, with oxygen therapy and mechanical ventilation to support injured organs. The exact components required to optimize resuscitation, such as choice and amount of fluid, appropriate type and intensity of hemodynamic monitoring, and role of adjunctive vasoactive agents, all remain controversial, even after the completion and reporting of recent large randomized trials.
Evidence from an older study suggests that protocol-based, early goal-directed therapy (EGDT) may confer a greater survival advantage than clinical assessments of organ perfusion and management without a protocol. EGDT included an aggressive resuscitation protocol with specific hemodynamic thresholds for fluid administration, blood transfusion, and use of ionotropes. Given the many controversial features of this older single-center trial, the recent ProCESS trial compared protocol-based standard care with protocol-based EGDT and usual care in >31 emergency departments in the United States. Among 1341 patients, the 60-day in-hospital mortality rate for protocol-based standard care (18.2%) was similar to that for usual care (18.9%) and protocol-based EGDT (21%). The ARISE trial confirmed this finding, showing that, among 1600 patients with early septic shock at 51 centers in Australia and New Zealand, 90-day mortality was similar for EGDT and usual care. Finally, the ProMISe trial, which enrolled 1260 patients in 56 hospitals in England, found that EGDT offered no mortality benefit in early septic shock but did increase treatment intensity and cost. Multiple subsequent meta-analyses of the ProCESS, ARISE, and ProMISe trials confirmed that EGDT offers no mortality benefit while increasing health care utilization and ICU admission in well-resourced countries. Modified versions of EGDT were also tested in lower-resourced settings, with no change in outcome. Thus EGDT is no longer recommended as the primary strategy for early resuscitation in septic shock. Nonetheless, some form of resuscitation is considered essential, and a standardized approach, akin to the use of “trauma teams,” has been advocated to ensure prompt care. The patient should be moved to an appropriate setting, such as the ICU, for ongoing care. SUBSEQUENT TREATMENT OF SEPSIS AND SEPTIC SHOCK
After initial resuscitation, attention is focused on monitoring and support of organ function, avoidance of complications, and de-escalation of care when possible. Monitoring
Hemodynamic monitoring devices may clarify the primary physiologic manifestations in sepsis and septic shock. The clinical usefulness of these monitoring devices can be attributable to the device itself, the algorithm linked to the device, or the static/dynamic target of the algorithm. Decades ago, the standard care of shock patients included invasive devices like the pulmonary artery catheter (PAC), also known as the continuous ScvO2 catheter. The PAC can estimate cardiac output and measure mixed venous oxygen saturation, among other parameters, to refine the etiology of shock and potentially influence patient outcomes. Recently, a Cochrane review of 2923 general-ICU patients (among whom the proportion of patients in shock was not reported) found no difference in mortality with or without PAC management, and the PAC therefore is no longer recommended for routine use. Instead, a variety of noninvasive monitoring tools, such as arterial pulse contour analysis (PCA) or focused echocardiography, can provide continuous estimates of parameters such as cardiac output, beat-to-beat stroke volume, and pulse pressure variation. These tools, along with passive leg-raise maneuvers or inferior vena cava collapsibility on ultrasound, can help determine a patient’s volume responsiveness but require that a variety of clinical conditions be met (e.g., patient on mechanical ventilation, sinus rhythm); in addition, more evidence from larger randomized trials on the impact of these tools in daily management is needed. Support Of Organ Function
The primary goal of organ support is to improve delivery of oxygen to the tissues as quickly as possible. Depending on the underlying physiologic disturbance, this step may require administration of IV fluids or vasopressors, blood transfusions, or ventilatory support.
Many crystalloids can be used in septic shock, including 0.9% normal saline, Ringer’s lactate, Hartmann’s solution, and Plasma-Lyte. Because crystalloid solutions vary in tonicity and inorganic/organic anions, few of these preparations closely resemble plasma. Normal saline is widely used in the United States. Colloid solutions (e.g., albumin, dextran, gelatins, or hydroxyethyl starch) are the most widely used fluids in critically ill patients, with variability across ICUs and countries. A clinician’s choice among colloids is influenced by availability, cost, and the desire to minimize interstitial edema. Many think that a greater intravascular volume is gained by use of colloids in shock, but the effects of colloids are modified by molecular weight and concentration as well as by vascular endothelial changes during inflammation. A network meta-analysis using direct and indirect comparisons in sepsis found evidence of higher mortality with starch than with crystalloids (relative risk [RR], 1.13; 95% CI, 0.99–1.30 [high confidence]) and no difference between albumin (RR, 0.83; 95% CI, 0.65, 1.04 [moderate confidence]) or gelatin (RR, 1.24; 95%CI, 0.61, 2.55 [very low confidence]) and crystalloids. In general, crystalloids are recommended on the basis of strong evidence as first-line fluids for sepsis resuscitation, with specific caveats; their use is guided by resolution of hypotension, oliguria, altered mentation, and hyperlactemia. Only weak evidence supports the use of balanced crystalloids, and guidelines recommend against using hydroxyethyl starches for intravascular volume replacement.
When circulating fluid volume is adequate, vasopressors are recommended to maintain perfusion of vital organs. Vasopressors such as norepinephrine, epinephrine, dopamine, and phenylephrine differ in terms of half-life, β- and α-adrenergic stimulation, and dosing regimens. Recent evidence comes from the SOAP II trial, a double-blind randomized clinical trial at eight centers comparing norepinephrine with dopamine in 1679 undifferentiated ICU patients with shock, of whom 63% were septic. Although no difference was observed in 28-day mortality or in predefined septic-shock subgroup, arrhythmias were significantly greater with dopamine. These findings were confirmed in a subsequent meta-analysis. As a result, expert opinion and consensus guidelines recommend norepinephrine as the first-choice vasopressor in septic shock. Levels of the endogenous hormone vasopressin may be low in septic shock, and the administration of vasopressin can reduce the norepinephrine dose. Consensus guidelines suggest adding vasopressin (up to 0.03 U/min) in patients without a contraindication to norepinephrine, with the intent of raising mean arterial pressure or decreasing the norepinephrine dose. There may be select indications for use of alternative vasopressors—e.g., when tachyarrhythmias from dopamine or norepinephrine, limb ischemia from vasopressin, or other adverse effects dictate.
The transfusion of red blood cells to high thresholds (>10 g/dL) had been suggested as part of EGDT in septic shock. However, the recent Scandinavian TRISS trial in 1005 septic shock patients demonstrated that a lower threshold (7 g/dL) resulted in 90-day mortality rates similar to those with a higher threshold (9 g/dL) and reduced transfusions by almost 50%.
Significant hypoxemia (PaO2, <60 mmHg; or SaO2, <90%), hypoventilation (rising PaCO2), increased work of breathing, and inadequate or unsustainable compensation for metabolic acidosis (pH <7.20) are common indications for mechanical ventilatory support. Endotracheal intubation protects the airway, and positive-pressure breathing allows oxygen delivery to metabolically active organs in favor of inspiratory muscles of breathing and the diaphragm. An experiment in dogs showed that the relative proportion of cardiac output delivered to respiratory muscles in endotoxic shock decreased by fourfold with spontaneous ventilation over that with mechanical ventilation. During intubation, patients in shock should be closely monitored for vasodilatory effects of sedating medications or compromised cardiac output due to increased intrathoracic pressure, both of which may cause hemodynamic collapse. With hemodynamic instability, noninvasive mask ventilation may be less suitable in patients experiencing sepsis-associated acute respiratory failure. Adjuncts
One of the great disappointments in sepsis management over the past 30 years has been the failure to convert advances in our understanding of the underlying biology into new therapies. Researchers have tested both highly specific agents and those with more pleotropic effects. The specific agents can be divided into those designed to interrupt the initial cytokine cascade (e.g., anti-LPS or anti-proinflammatory cytokine strategies) and those that interfere with dysregulated coagulation (e.g., antithrombin or activated protein C). Recombinant activated protein C (aPC) was one of the first agents approved by the U.S. Food and Drug Administration and was the most widely used. A large, randomized, double-blind, placebo-controlled, multicenter trial of aPC in severe sepsis (the PROWESS trial) was reported in 2001; the data suggested an absolute risk reduction of up to 6% among aPC-treated patients with severe sepsis. However, subsequent phase 3 trials failed to confirm this effect, and the drug was withdrawn from the market. It is no longer recommended in the care of sepsis or septic shock.
Many adjunctive treatments in sepsis and septic shock target changes in the innate immune response and coagulation cascade. Specific adjuncts like glucocorticoids in septic shock have continued to be widely used. A large negative clinical trial and a conflicting systematic review in 2009 extended the debate about whether glucocorticoids lower 28-day mortality or improve shock reversal. Most meta-analyses report no change in mortality but an increase in shock reversal with glucocorticoid treatment. The recent HYPRESS trial found no difference between patients with severe sepsis who were treated with glucocorticoids and control patients in terms of the development of shock or the mortality rate. These data and others led to a suggestion in international clinical practice guidelines against using IV hydrocortisone to treat septic shock if adequate fluid resuscitation and vasopressor therapy are able to restore hemodynamic stability. If not, the guidelines suggest the administration of IV hydrocortisone at a dose of 200 mg per day (weak recommendation, low quality of evidence).
Among other adjuncts, IV immunoglobulin may be associated with potential benefit, but significant questions remain and such treatment is not part of routine practice. Despite a large number of observational studies suggesting that statin use mitigates the incidence or outcome of sepsis and severe infection, there are no confirmatory randomized controlled trials, and statins are not an element in routine sepsis care. De-Escalation of Care
Once patients with sepsis and septic shock are stabilized, it is important to consider which therapies are no longer required and how care can be minimized. The de-escalation of initial broad-spectrum therapy, which observational evidence indicates is safe, may reduce the emergence of resistant organisms as well as potential drug toxicity and costs. The added value of combination antimicrobial therapy over that of adequate single-agent antibiotic therapy in severe sepsis has not been established. Current guidelines recommend combination antimicrobial therapy only for neutropenic sepsis and sepsis caused by Pseudomonas. Large trials are under way in the United States to determine how serum biomarkers like procalcitonin can assist clinicians in minimizing antibiotic exposure, while European trials are indicating that this biomarker may lead to a reduction in the duration of treatment and in daily defined doses in critically ill patients with a presumed bacterial infection.
TABLE 297-2Elements of Care in Sepsis and Septic Shock: Recommendations Adapted from International Consensus Guidelines ||Download (.pdf) TABLE 297-2 Elements of Care in Sepsis and Septic Shock: Recommendations Adapted from International Consensus Guidelines
|Sepsis and septic shock constitute an emergency, and treatment should begin right away. |
|Resuscitation with IV crystalloid fluid (30 mL/kg) should begin within the first 3 h. |
|Saline or balanced crystalloids are suggested for resuscitation. |
|If the clinical examination does not clearly identify the diagnosis, hemodynamic assessments (e.g., with focused cardiac ultrasound) can be considered. |
|In patients with elevated serum lactate levels, resuscitation should be guided towards normalizing these levels when possible. |
|In patients with septic shock requiring vasopressors, the recommended target mean arterial pressure is 65 mmHg. |
|Hydroxyethyl starches and gelatins are not recommended. |
|Norepinephrine is recommended as the first-choice vasopressor. |
|Vasopressin should be used with the intent of reducing the norepinephrine dose. |
|The use of dopamine should be avoided except in specific situations—e.g., in those patients at highest risk of tachyarrhythmias or relative bradycardia. |
|Dobutamine use is suggested when patients show persistent evidence of hypoperfusion despite adequate fluid loading and use of vasopressors. |
|Red blood cell transfusion is recommended only when the hemoglobin concentration decreases to <7.0 g/dL in the absence of acute myocardial infarction, severe hypoxemia, or acute hemorrhage. |
|Infection Control |
|So long as no substantial delay is incurred, appropriate samples for microbiologic cultures should be obtained before antimicrobial therapy is started. |
|IV antibiotics should be initiated as soon as possible (within 1 h); specifically, empirical broad-spectrum therapy should be used to cover all likely pathogens. |
|Antibiotic therapy should be narrowed once pathogens are identified and their sensitivities determined and/or once clinical improvement is evident. |
|If needed, source control should be undertaken as soon as is medically and logistically possible. |
|Daily assessment for de-esclation of antimicrobial therapy should be conducted. |
|Respiratory Support |
|A target tidal volume of 6 mL/kg of predicted body weight (compared with 12 mL/kg in adult patients) is recommended in sepsis-induced ARDS. |
|A higher PEEP rather than a lower PEEP is used in moderate to severe sepsis-induced ARDS. |
|In severe ARDS (PaO2/FIO2, <150 mmHg), prone positioning is recommended, and recruitment maneuvers and/or neuromuscular blocking agents for ≤48 h are suggested. |
|A conservative fluid strategy should be used in sepsis-induced ARDS if there is no evidence of tissue hypoperfusion. |
|Routine use of a pulmonary artery catheter is not recommended. |
|Spontaneous breathing trials should be used in mechanically ventilated patients who are ready for weaning. |
|General Supportive Care |
|Patients requiring a vasopressor should have an arterial catheter placed as soon as is practical. |
|Hydrocortisone is not suggested in septic shock if adequate fluids and vasopressor therapy can restore hemodynamic stability. |
|Continuous or intermittent sedation should be minimized in mechanically ventilated sepsis patients, with titration targets used whenever possible. |
|A protocol-based approach to blood glucose management should be used in ICU patients with sepsis, with insulin dosing initiated when two consecutive blood glucose levels are >180 mg/dL. |
|Continuous or intermittent renal replacement therapy should be used in patients with sepsis and acute kidney injury. |
|Pharmacologic prophylaxis (unfractionated heparin or low-molecular-weight heparin) against venous thromboembolism should be used in the absence of contraindications. |
|Stress ulcer prophylaxis should be given to patients with risk factors for gastrointestinal bleeding. |
|The goals of care and prognosis should be discussed with patients and their families. |
TABLE 297-3Initial Antimicrobial Therapy for Severe Sepsis with No Obvious Source in Adults with Normal Renal Function ||Download (.pdf) TABLE 297-3 Initial Antimicrobial Therapy for Severe Sepsis with No Obvious Source in Adults with Normal Renal Function
Septic shock (immunocompetent adult)
The many acceptable regimens include (1) piperacillin-tazobactam (3.375–4.5 g q6h), (2) cefepime (2 g q12h), or (3) meropenem (1 g q8h) or imipenem-cilastatin (0.5 g q6h). If the patient is allergic to β-lactam antibiotics, use (1) aztreonam (2 g q8h) or (2) ciprofloxacin (400 mg q12h) or levofloxacin (750 mg q24h). Add vancomycin (loading dose of 25–30 mg/kg, then 15–20 mg/kg q8–12h) to each of the above regimens.
Neutropenia (<500 neutrophils/μL)
Regimens include (1) cefepime (2 g q8h), (2) meropenem (1 g q8h) or imipenem-cilastatin (0.5 g q6h) or doripenem (500 mg q8h), or (3) piperacillin-tazobactam (3.375 g q4h). Add vancomycin (as above) if the patient has a suspected central line–associated bloodstream infection, severe mucositis, skin/soft tissue infection, or hypotension. Add tobramycin (5–7 mg/kg q24h) plus vancomycin (as above) plus caspofungin (one dose of 70 mg, then 50 mg q24h) if the patient has severe sepsis/septic shock.
Use ceftriaxone (2 g q24h, or—in meningitis—2 g q12h). If the local prevalence of cephalosporin-resistant pneumococci is high, add vancomycin (as above). If the patient is allergic to β-lactam antibiotics, use levofloxacin (750 mg q24h) or moxifloxacin (400 mg q24h) plus vancomycin (as above).
Before modern intensive care, sepsis and septic shock were highly lethal, with infection leading to compromise of vital organs. Even with intensive care, nosocomial mortality rates for septic shock often exceeded 80% as recently as 30 years ago. Now, the U.S. Burden of Disease Collaborators report that the primary risk factor for sepsis and septic shock—i.e., infection—is the fifth leading cause of years of productive life lost because of premature death. More than half of sepsis cases require ICU admission, representing 10% of all ICU admissions. However, with advances in training, surveillance, monitoring, and prompt initiation of supportive care for organ dysfunction, the mortality rate from sepsis and septic shock is now closer to 20% in many series. Although some data suggest that mortality trends are even lower, attention has been focused on the trajectory of recovery among survivors. Patients who survive to hospital discharge after sepsis remain at increased risk of death in the following months and years. Those who survive often suffer from impaired physical or neurocognitive dysfunction, mood disorders, and low quality of life. In many studies, it is difficult to determine the causal role of sepsis. However, an analysis of the Health and Retirement Study—a large longitudinal cohort study of aging Americans—suggested that severe sepsis significantly accelerated physical and neurocognitive decline. Among survivors, the rate of hospital readmission within 90 days after sepsis exceeds 40%.
In light of the persistently high mortality risk in sepsis and septic shock, prevention may be the best approach to reducing avoidable deaths, but preventing sepsis is a challenge. The aging of the population, the overuse of inappropriate antibiotics, the rising incidence of resistant microorganisms, and the use of indwelling devices and catheters contribute to a steady burden of sepsis cases. The number of cases could be reduced by avoiding unnecessary antibiotic use, limiting use of indwelling devices and catheters, minimizing immune suppression when it is not needed, and increasing adherence to infection control programs at hospitals and clinics. To facilitate earlier treatment, such pragmatic work could be complemented by research into the earliest pathophysiology of infection, even when symptoms of sepsis are nascent. In parallel, the field of implementation science could inform how best to increase adoption of infection control in high-risk settings and could guide appropriate care.
et al: Epidemiology of severe sepsis in the United States: Analysis of incidence, outcome, and associated costs of care. Crit Care Med 29:1303, 2001.
et al: Immunosuppression in patients who die of sepsis and multiple organ failure. JAMA 306:2594, 2011.
et al: Assessment of global incidence and mortality of hospital-treated sepsis. Current estimates and limitations. Am J Respir Crit Care Med 193:259, 2016.
et al: Disease tolerance as a defense strategy. Science 335:936, 2012.
et al: Surviving sepsis campaign: International guidelines for management of sepsis and septic shock: 2016. Crit Care Med 45:486, 2017.
et al: Fluid resuscitation in sepsis: A systematic review and network meta-analysis. Ann Intern Med 161:347, 2014.
et al: Assessment of clinical criteria for sepsis: For the Third International Consensus Definitions for Sepsis and Septic Shock (Sepsis-3). JAMA 315:762, 2016.
et al: The SOFA (sepsis-related organ failure assessment) score to describe organ dysfunction/failure. On behalf of the Working Group on Sepsis-Related Problems of the European Society of Intensive Care Medicine. Intensive Care Med 22:707, 1996.
et al: A randomized trial of protocol-based care for early septic shock. N Engl J Med 370:1683, 2014.
Additional Online References
et al: A framework for the development and interpretation of different sepsis definitions and clinical criteria. Crit Care Med 44:e113, 2016.
et al: Corticosteroids in the treatment of severe sepsis and septic shock in adults: A systematic review. JAMA 301:2362, 2009.
et al: Efficacy and safety of recombinant human activated protein C for severe sepsis. N Engl J Med 344:699, 2001.
et al: The ACCP-SCCM consensus conference on sepsis and organ failure. Chest 101:1481, 1992.
et al: Non-invasive ventilation in community-acquired pneumonia and severe acute respiratory failure. Intensive Care Med 38:458, 2012.
et al: Alarmins: Awaiting a clinical response. J Clin Invest 122:2711, 2012.
USBoD: The state of US health, 1990–2010: Burden of diseases, injuries, and risk factors. JAMA 310:591, 2013.
et al: Efficacy and safety of procalcitonin guidance in reducing the duration of antibiotic treatment in critically ill patients: A randomised, controlled, open-label trial. Lancet Infect Dis 16:819, 2016.
et al: Paresis acquired in the intensive care unit: A prospective multicenter study. JAMA 288:2859, 2002.
et al: Improvement in process of care and outcome after a multicenter severe sepsis educational program in Spain. JAMA 299:2294, 2008.
et al: Has the mortality of septic shock changed with time? Crit Care Med 26:2078, 1998.
et al: Sepsis and septic shock: A history. Crit Care Clin 25:83, 2009.
et al: Lower versus higher hemoglobin threshold for transfusion in septic shock. N Engl J Med 371:1381, 2014.
et al: Sepsis-induced apoptosis causes progressive profound depletion of B and CD4+ T lymphocytes in humans. J Immunol 166:6952, 2001.
et al: Depletion of dendritic cells, but not macrophages, in patients with sepsis. J Immunol 168:2493, 2002.
C: Distribution of respiratory muscle and organ blood flow during endotoxic shock in dogs. J Appl Physiol 59:1802, 1985.
et al: Long-term cognitive impairment and functional disability among survivors of severe sepsis. JAMA 304:1787, 2010.
et al: Systemic inflammatory response syndrome criteria for severe sepsis. N Engl J Med 373:881, 2015.
et al: Mortality related to severe sepsis and septic shock among critically ill patients in Australia and New Zealand, 2000–2012. JAMA 311:1308, 2014.
et al: Effect of hydrocortisone
on development of shock among patients with severe sepsis: The HYPRESS randomized clinical trial. JAMA 316:1775, 2016.
et al: Duration of hypotension before initiation of effective antimicrobial therapy is the critical determinant of survival in human septic shock. Crit Care Med 34:1589, 2006.
et al: Recommendations for the diagnosis and management of corticosteroid insufficiency in critically ill adult patients: Consensus statements from an international taskforce by the American College of Critical Care Medicine. Crit Care Med 36:1937, 2008.
et al: Infection rate and acute organ dysfunction risk as explanations for racial differences in severe sepsis. JAMA 303:2495, 2010.
et al: Trial of early, goal-directed resuscitation for septic shock. N Engl J Med 373:577, 2015.
MG: Resuscitation fluids. N Engl J Med 369:2462, 2013.
RA: Systemic steroids in severe sepsis and septic shock. Am J Respir Crit Care Med 185:133, 2012.
et al: Goal-directed resuscitation for patients with early septic shock. N Engl J Med 371:1496, 2014.
et al: Readmission diagnoses after hospitalization for severe sepsis and other acute medical conditions. JAMA 313:1055, 2015.
et al: Acute respiratory distress syndrome: The Berlin definition. JAMA 307:2526, 2012.
et al: Drotrecogin alfa (activated) in adults with septic shock. N Engl J Med 366:2055, 2012.
et al: New approaches to sepsis: Molecular diagnostics and biomarkers. Clin Microbiol Rev 25:609, 2012.
et al: Objective sepsis surveillance using electronic clinical data. Infect Control Hosp Epidemiol 37:163, 2016.
et al: Early goal-directed therapy in the treatment of severe sepsis and septic shock. N Engl J Med 345:1368, 2001.
MR: Septic shock: Advances in diagnosis and treatment. JAMA 314:708, 2015.
et al: Application of a framework to assess the usefulness of alternative sepsis criteria. Crit Care Med 44:e122, 2016.
et al: The Third International Consensus Definitions for Sepsis and Septic Shock (Sepsis-3). JAMA 315:801, 2016.
et al: The impact of timing of antibiotics on outcomes in severe sepsis and septic shock: A systematic review and meta-analysis. Crit Care Med 43:1907, 2015.
S: Pattern recognition receptors and inflammation. Cell 140:805, 2010.
et al: Macroscopic postmortem findings in 235 surgical intensive care patients with sepsis. Anesth Analg 108:1841, 2009.
Van der Poll
SM: Host–pathogen interactions in sepsis. Lancet Infect Dis 8:32, 2008.
et al: Incidence, risk factors, and attributable mortality of secondary infections in the intensive care unit after admission for sepsis. JAMA 315:1469, 2016.
JL, De Backer
D: Circulatory shock. N Engl J Med 369:1726, 2013.
et al: International study of the prevalence and outcomes of infection in intensive care units. JAMA 302:2323, 2009.
LDH: Shock, in Principles of Critical Care. J
Wood (eds). New York, McGraw-Hill, 1998, pp 277–301.
MA: The acute respiratory distress syndrome. N Engl J Med 342:1334, 2000.
et al: Sepsis-induced acute kidney injury revisited: Pathophysiology, prevention and future therapies. Curr Opin Crit Care 20:588, 2014.