Chapter 138. Sepsis

1. What is the definition of systemic inflammatory response syndrome (SIRS) and how do you differentiate SIRS from sepsis, severe sepsis, and septic shock?

2. Which patients presenting with sepsis need admission to the intensive care unit (ICU)?

3. Which septic patients need invasive monitoring (arterial catheter, central venous catheter)?

4. What interventions in the treatment of sepsis improve mortality?

5. Which septic patients deserve empiric steroids as part of the therapeutic regimen?

Sepsis is a clinical syndrome that complicates severe infection and is characterized by systemic inflammation and widespread tissue injury. The incidence and number of sepsis-related deaths has increased from 1979 to 2000, and sepsis remains the tenth most common cause of death in the United States. Despite the rising number of cases, earlier identification of sepsis and improved intensive medical care has been shown to reduce the overall mortality rate to approximately 17.9%. Severity is correlated with mortality; up to 40% of patients with severe sepsis and 60% of patients with septic shock die due to sepsis-related complications (Table 138-1).

Successful shock resuscitation may be associated with considerable morbidity and mortality. Multiple organ dysfunction syndrome (MODS) refers to severe acquired dysfunction of at least two organ systems lasting at least 24 to 48 hours in the setting of sepsis, trauma, burns, or severe inflammatory conditions so that homeostasis cannot be maintained without intervention. Mortality is directly correlated with the number of dysfunctional organs and the duration of dysfunction (Table 138-2). An uncontrolled hyperinflammatory response is believed to be the cause of multiple organ dysfunction.

The American College of Chest Physicians (ACCP) and the Society of Critical Care Medicine (SCCM) defined in 2001 at the International Sepsis Definitions Conference the following terms to describe the spectrum of systemic inflammation and sepsis. Systemic inflammatory response syndrome (SIRS) is defined as a clinical syndrome that results from activation of the immune system whether due to infection, trauma, burns, or a noninfectious inflammatory process. It includes at least two of the following: (1) temperature > 38 degrees Celsius or < 36 degrees Celsius; (2) heart rate > 90 beats/minute; (3) respiratory rate > 20 breaths/minute or paCO2 < 32 mm Hg; a (4) white blood cell count > 12,000 cells/mm3, or < 4000 cells/mm3, or > 10% immature (band) forms.

Sepsis is defined as a clinical syndrome that results from activation of the immune system with a documented infection. The definition of sepsis includes the above SIRS criteria plus a culture-proven infection or presumed presence of an infection. The severity of sepsis is graded according to the associated organ dysfunction and hemodynamic compromise. Severe sepsis is defined as the presence of sepsis and one or more organ dysfunctions. Organ dysfunction can be defined as acute lung injury including acute respiratory distress syndrome (ARDS), disseminated intravascular coagulation, thrombocytopenia, altered mental status, mottled skin, capillary refill greater than three seconds, renal dysfunction, hepatic dysfunction, cardiac dysfunction based on echocardiography or measurement of cardiac index, or lactic acidosis indicating hypoperfusion. Septic shock is defined as the presence of sepsis and refractory hypotension with mean systemic blood pressure lower than 65 mm Hg unresponsive to crystalloid fluid challenge of 20 to 40 cc/kg leading to acute circulatory collapse.

Sepsis is as an uncontrolled inflammatory response to an infection in which a dysregulated host immune response leads to multiorgan involvement not limited to the source infected organ. Microbial antigens such as lipopolysaccharides (LPS) from Gram-negative bacteria bind to Toll-like receptors on inflammatory cells, thereby causing a complex immune reaction involving T-cells, macrophages, neutrophile, endothelial cells, and dendritic cells. Cytokines (such as IL-1, IL-6, IL-8), growth factors (such as TNFa), high-mobility group box-1 (HMGB-1), arachidonic acid metabolites, and nitric oxide and host genetics likely determine the nature of the response. The complement cascade, coagulation cascades, platelets, and leukocytes interact at the vascular endothelium level resulting in microvascular injury, thrombosis, and loss of endothelial integrity, which altogether results in tissue ischemia. This diffuse endothelial disruption is responsible for the various organ dysfunctions and global tissue hypoxia that accompany severe sepsis and septic shock. Global tissue hypoxia can occur from multiple mechanisms including decreased preload, vasoregulatory dysfunction, myocardial depression, and impaired tissue extraction due to microcirculatory dysfunction or mitochondrial dysfunction (cytopathic hypoxia). Some noninfectious processes (eg, pancreatitis) can also lead to a dysregulated host immune response and multiorgan dysfunction, and these conditions are categorized using the term SIRS. These patients appear septic without a clear infectious source.

The differential diagnosis for sepsis is broad, including conditions that present with high-output states and those with low-output shock. While many patients present with common conditions that meet SIRS criteria—nonmassive pulmonary embolus, alcohol withdrawal, even COPD exacerbations—other conditions can mimic sepsis with high cardiac output state and wide pulse pressure without shock including thyrotoxicosis, aortic regurgitation, arteriosclerosis, and cirrhosis. While all patients with sepsis do not present in shock, the differential diagnosis for shock should always be considered. In addition to septic shock, conditions that belong to the category of vasodilatory, or high cardiac output, shock include anaphylaxis, adrenal insufficiency, and neurogenic shock.

The other causes of shock should always remain in the differential for patients with sepsis as the presentation may combine two or more processes or may simply mimic two processes. For example, a patient with underlying poor cardiac function who becomes septic may not be able to generate the high cardiac outputs expected in sepsis and may not have the typical findings on physical examination of a septic patient. Moreover, the phenomenon of sepsis-induced myocardial dysfunction occurs when patients have normal cardiac function prior to their infection and the sepsis induces a global cardiac dysfunction, seen on echocardiography as global hypokinesis, which may impair the expected high cardiac output usually associated with a vasodilated circulation. The other causes of shock all fall into a category of low-output states, including cardiogenic shock, hypovolemic shock, and obstructive shock (Table 138-3).

Diagnosis

A high index of suspicion for the clinical sepsis syndrome facilitates early identification of the inciting stimulus, and intervention may limit the ongoing immune response. Septic patients present with signs and symptoms ranging from fever and chills to delirium to shock. Presentation often varies according to infection source, patient age, underlying comorbidities (including immune system function and cardiac status), and timing of presentation relative to onset of sepsis. Early manifestations of sepsis—tachycardia, oliguria, and hyperglycemia—may be subtle and easily overlooked in the hospitalized patient. Signs of established sepsis include altered mental status, metabolic acidosis and respiratory alkalosis, hypotension with decreased systemic vascular resistance (SVR) and elevated cardiac output, and coagulopathy. Late manifestations include acute lung injury (ALI), ARDS, acute renal failure, hepatic dysfunction, and refractory shock. Sepsis can occur related to a systemic inflammatory response to any infectious source, but less than 50% of septic patients will have positive blood cultures. Therefore, aggressive clinical evaluation should occur inquiring about symptoms including fever, rigors, mental status changes, cough, shortness of breath, vomiting, diarrhea, dysuria, hematuria, and many other symptoms that might suggest the source of infection. Patients may also present with obvious or subtle physical examination findings. In an analogous manner to the thorough history, a complete physical examination can assess for potential infection sources from any organ system. Sometimes inconspicuous and missed infection sources may include the skin and soft tissue, central nervous system, gastrointestinal tract, and indwelling devices.

In addition to stabilizing the patient, it is critical to identify the cause of the ongoing immunologic response. Obtaining cultures for blood, urine, and other fluids early, prior to administration of antibiotics, helps preserve the integrity of results and should be an extremely high priority, but not at the expense of administering antibiotics expediently. Identification of the underlying source remains paramount, and lack of source identification and control may render choice of antibiotics meaningless. The most common sites of infection in sepsis are the urinary and respiratory tracts, but any organ system may be involved. Urinary sources include cystitis, pyelonephritis, and perinephric abscess. Patients with kidney stones may develop Gram-negative septicemia. Sinusitis, mastoiditis, pneumonia, lung abscess, and empyema may be associated with sepsis. Gastrointestinal sources of sepsis may include esophageal rupture or perforation following a procedure or after vomiting, cholangitis, cholecystitis, intestinal infarction or perforation, acute pancreatitis, Clostridium dificile colitis, diverticulitis, and intraabdominal abscess. Postoperative mediastinitis and acute bacterial endocarditis may lead to sepsis. It is also critical not to consider skin and soft tissue sources of sepsis because of the possibility of an infected decubitus ulcer, postoperative wound, soft tissue abscess, or necrotizing fasciitis. Vascular causes include catheter infections (central and peripheral lines, arterial catheters, dialysis catheters, ventriculoperitoneal shunts) and septic thrombophlebitis. Infected articular prosthetic devices have also been associated with sepsis. Meningitis and intracranial abscess, sometimes associated with neurosurgery, are also considerations.

Relevant diagnostic studies based on symptoms, signs, and clinical suspicion in patients may include chest or abdominal radiography and culture of blood, urine, sputum, or other relevant body fluids that may be infected—cerebrospinal fluid (CSF) analysis, paracentesis in patients with ascites, or thoracentesis in patients with pleural effusion. Bacteremia is found in only approximately 50% of cases of severe sepsis and septic shock, while 20% to 30% of patients will have no microbial cause identified from any source. When relatively basic diagnostic evaluations—including plain films, blood cultures, and fluid cultures—do not yield a likely infectious culprit, advanced imaging including chest and abdominal computed tomography can assess for pulmonary infiltrates, intraabdominal abscesses, and obstructing renal stones. Biliary pathology may be better imaged with ultrasound, magnetic resonance imaging (MRI), or endoscopic retrograde pancreatography (ERCP) if necessary. Echocardiography to assess for endocarditis may be necessary based on physical examination findings, blood culture results, and the clinical picture.

All patients with a presentation of severe sepsis or septic shock should be admitted to or transferred to a monitored setting that is capable of continuous vital sign monitoring with the ability to measure central venous pressure (CVP) and central venous oxygen saturations (ScvO2). Those patients with SIRS and sepsis should be monitored closely if not placed in an intensive care unit setting so that they can be treated promptly if they start to show signs of deterioration. Vital signs should be monitored frequently and a low threshold to utilize other monitoring (telemetry and continuous pulse oximetry) should be considered. Patient response to initial resuscitation in the emergency department or on a medical floor should guide the clinician with respect to intensive care unit (ICU) admission with the capability of intraarterial blood pressure measurement and the delivery of vasoactive medications. Intermediate care units (sometimes called step-down units or transitional care units) vary from facility to facility in their capabilities for invasive monitoring and use of vasoactive agents. The protocols and policies at individual institutions will help determine placement of these patients, based on monitoring requirements. Due to the emphasis placed on early aggressive resuscitation, early antibiotics, and early source identification and control, septic patients require careful evaluation to determine admission location, and patients with marginally stable clinical parameters should be admitted to an ICU setting to expeditiously meet early care goals to improve outcomes.

Management of severe sepsis and septic shock requires a structured approach that assures proper diagnostic evaluation and implementation of evidence-based interventions in an expedient manner to improve outcomes (Figure 138-1). This approach requires (1) empiric antibiotic coverage of an infectious source while cultures are pending, (2) optimal fluid resuscitation, (3) pressor and/or inotrope therapy for selected patients, and (4) consideration of additional therapies such as drainage of abscesses, removal of lines, moderate (but not intensive) control of hyperglycemia, and (5) consideration of steroids in selected patient subsets when indicated.

Antibiotic Therapy

Initial emergency department management of patients with sepsis begins with a high awareness of the condition by assessing all patients for the presence of SIRS criteria. Numerous studies have shown that early and appropriate antibiotics are associated with markedly improved outcomes. Antibiotics should be directed against the likely organisms based upon the presumptive infection source. In many situations, especially when the presumptive source of infection is not obvious, multiple antibiotic agents should be initiated to offer broad antimicrobial coverage. Such broad coverage should then be reevaluated daily to optimize dosing and minimize drug interactions and the development of resistance. Choice of antibiotic depends upon penetration into the suspected infection site, local resistance patterns, efficacy against the most likely organisms, and risks of side effects. Therapeutic drainage of an infected space is critical to diagnose the source of infection, guide the choice of antibiotic therapy, and facilitate recovery. In patients with devices, clinicians may need to evaluate and consider early and rapid removal of potentially or known infected invasive devices including central venous catheters (CVCs), peripherally inserted central venous catheters (PICCs), foley catheters, and other implanted hardware.

Recent evidence suggests that mortality increases with delay of antibiotics more than one hour after identification and management of severe sepsis or septic shock. In selected patients, including abdominal surgery patients and patients receiving total parenteral nutrition (TPN) or long-standing steroids, the suspicion for fungal infections should be raised and may require adding empiric antifungal agents to the antimicrobial regimen.

Intravenous Fluids

Volume resuscitation should begin simultaneously with empiric antibiotic therapy in patients suspected of having sepsis. In the vasodilatory state low blood pressures with decreased venous return lead to an underfilled, but hyperdynamic heart. Rivers and colleagues showed that early goal-directed therapy (EGDT), initiated in the emergency department, improved mortality in patients with severe sepsis and septic shock. That early goal-directed therapy includes early aggressive volume resuscitation in the first six hours of care, and other measures over the first hours and days of care (see Figure 138-1). As part of EGDT, a central venous catheter placed in the internal jugular or subclavian vein allows close monitoring of central venous pressure (CVP). If the patient remains hypotensive (mean arterial pressure [MAP] < 65mm Hg) after a reasonable fluid challenge with crystalloid (approximately 20–40 cc/kg) to optimize filling pressures, CVP and ScvO2 monitoring allows adjustment of or addition of interventions based on the parameters measured within the individual patient to achieve the goal of ScvO2 at 70%. Studies evaluating resuscitation with crystalloid intravenous (IV) fluids versus colloid fluid (eg, albumin) support no added benefit of colloid fluid over crystalloid. The Saline Versus Albumin Fluid Evaluation (SAFE) study evaluated nearly 7000 critically ill patients, 18% of whom had severe sepsis. Patients were randomly assigned to receive 4% albumin versus normal saline, and investigators reported no differences in mortality at 28 days. Additionally, there were no significant differences seen in the sepsis subgroup. Despite a theoretic benefit to using albumin in highly selected patients with significant volume overload, no studies support this approach. For routine use in sepsis, crystalloid fluid (normal saline) should be used first due to evidence of benefit, markedly lower expense, and demonstrated safety (lacking the inherent risks of blood product administration with albumin).

The EGDT algorithm uses a CVP goal of 8–12 cm H2O, which is a reasonable estimate goal. However, that goal should not be applied blindly to all patients without knowledge of coexisting conditions including pulmonary arterial hypertension, dilated cardiomyopathy, and old right ventricular infarction. Clinicians should follow the trend of the CVP and correlate it with the ScvO2, patient hemodynamics, and evidence of organ perfusion including mental status and urine output. Ample data suggests that the CVP serves as a poor predictor of volume responsiveness, and multiple factors are necessary to determine the need for continued volume resuscitation. The ScvO2 carries valuable information and weight, offering the clinician an assessment of cardiac function and oxygen delivery balanced against oxygen consumption.

Blood Transfusion

Patients with severe sepsis or septic shock who have been resuscitated adequately will usually demonstrate the physiology of a low oxygen extraction state with high ScvO2 values, but importantly, preresuscitation values may make patients appear as high oxygen extractors, with low ScvO2 values more consistent with a low cardiac output state. EGDT protocol includes transfusing red blood cells if the hematocrit is less than 30% and the ScvO2 remains less than 70% after meeting CVP and blood pressure goals. While a subgroup analysis favored transfusions to improve outcomes, potential deleterious effects from red blood cell transfusions, including questionable efficacy of older stored blood, the immunomodulating effects of red blood cell transfusions, and the risk of transfusion reactions, make this part of the EGDT protocol more difficult to recommend broadly for every patient meeting EGDT criteria. Routine transfusion of red blood cells should likely be avoided unless patients exhibit evidence of active bleeding, active ischemia, or hemologlobin is below 7 gm/dL. Expert opinion is conflicting with regard to blood transfusion threshold in sepsis management because other evidence-based literature supports no improvement in ICU patients who are transfused to a goal hemoglobin of 10 gm/dL rather than 7 gm/dL.

Vasoactive Medications

An important aspect of sepsis management includes vasoactive medications. Pressors are often needed to maintain mean arterial blood pressures (MAP) and the choice of agent depends on the physiologic need (Table 138-4). The EGDT protocol recommends vasopressor agents to maintain MAP ≥ 65 mmHg. There is no firm evidence favoring one pressor agent over another, but norepinephrine likely has the greatest vasoconstrictor potency along with some inotropic effect. Dopamine can be used if a more inotropic and chronotropic effect is desired and should be avoided if cardiogenic shock is suspected due to demonstrated increased mortality and arrhythmias in that patient population. Low-dose dopamine does not provide renal protection, and should not be used solely for that purpose. Phenylephrine may be the preferred agent for blood pressure elevation in patients with prohibitive tachycardia or arrhythmias. Vasopressin, a pure vasoconstrictor, has been used to lessen the doses of adrenergic pressor agents; however, current available data does not support its routine use in sepsis.

The reversible global myocardial dysfunction that often occurs during sepsis may manifest as poor organ perfusion or as persistently low ScvO2 values despite seemingly adequate filling pressures. Echocardiography or a pulmonary artery (PA) catheter may aid in this diagnosis; however, a PA catheter is rarely necessary in the management of septic shock. The addition of an inotropic agent such as dobutamine or milrinone may be temporarily required until the myocardial dysfunction improves. The EGDT protocol recommends inotropic agents to meet a goal ScvO2 of 70%—after achieving CVP, MAP, and hematocrit goals—in order to improve tissue oxygen delivery.

Nutrition

Septic patients typically have increased catabolic physiology (breakdown of protein in order to generate amino acids for gluconeogensis) due to inflammatory response. Supplying calories early in the clinical course of sepsis can limit catabolic activity; gut integrity and immune function may be better maintained, leading to decreased infections in critically ill patients. In general, gastrointestinal (GI) tract (enteral) feeding should begin when patients are unable to feed themselves or meet nutritional goals on their own. Several studies have shown a decrease in infectious complications and stress ulcer complications in patients who are started on early (<48 hours) enteral nutrition. Multiple factors, including existing nutritional status as well as the potential for harm from temporarily withholding enteral nutrition or starting parenteral nutrition, play into the decision of initiating enteral nutrition. Patients who are hemodynamically unstable or on high-dose vasporessors may bear risk of bowel ischemia if fed enterally. Total parenteral nutrition (TPN) should be reserved for a minority of patients, typically patients post gastrointestinal surgery. TPN leads to more infectious complications (including fungemia) than enteral feeding, though there is no data showing that enteral feeding improves mortality versus TPN.

Glycemic Control

Guidelines recommend moderate glycemic control (blood glucose [BG] 140–180 mg/dL) in medical ICU patients, including septic patients. Tight glycemic control (keeping BG 80–110 mg/dL) in critically ill medical patients leads to worse mortality outcomes. No significant mortality benefit is achieved with tight glycemic control in critically ill patients except for those undergoing cardiothoracic surgery. Since tight glycemic control (80–110 mg/dL) is associated with significant morbidity due to hypoglycemia, endocrine and critical care authorities recommend more moderate glycemic control (140–180 mg/dL) protocols in medical ICUs and for septic patients.

Steroids

Relative or functional adrenal insufficiency refers to suboptimal adrenal response to severe sepsis. While some debate has focused around when to prescribe corticosteroids to septic patients and how to make that decision, recent guidelines offer some assistance to clinicians making this decision (Table 138-5). In patients with septic shock, defined by systolic blood pressure less than 90 mm Hg despite volume resuscitation and vasopressor medications, corticosteroids offer benefit. The duration of the corticosteroids should be less than seven days and they should not be given with mineralocorticoid. Since the most recent guidelines by the Society of Critical Care Medicine (SCCM) in 2008, a 2009 meta-analysis summary of randomized controlled trial data suggests that low-dose (<300 mg hydrocortisone per day) short-course (<5 days) therapy offers significant mortality benefit (relative risk reduction 16%, number needed to treat 14) and length-of-stay reduction (average reduction 4.5 days) to patients with severe sepsis or septic shock.

Activated Protein C

Sepsis induces a procoagulant response and this led to the development of recombinant human activated protein C (APC), a protein that promotes fibrinolysis and inhibits production of thrombin. APC improves mortality in patients with severe sepsis or septic shock who have the highest baseline risk (defined as an APACHE II score > 25 or multiple organ dysfunction); patients with lower risk (APACHE II score < 25 or single organ dysfunction) receive no mortality benefit. In addition, APC may increase the risk of bleeding for all patient populations and extreme caution should be taken in the following patients: (1) severely thrombocytopenic (platelet count < 30,000/mm3), (2) conditions that increase risk of bleeding (eg, recent surgery requiring general or spinal anesthesia within 12 hours of infusion, intracranial surgery, recent stroke, cerebral aneurysm, recent gastrointestinal bleed), (3) recent anticoagulant therapy including unfractionated heparin, low-molecular-weight heparin, Coumadin, thrombolytics, glycoprotein IIb/IIIa antagonists), (4) portal hypertension. These patients were all excluded from the original study on APC. The decision to prescribe APC should be individualized (though limited to patients with severe sepsis or septic shock with high mortality risk) with respect to the patient's goals, values, and coexisting medical conditions including terminal illnesses. Because this medication bears high cost and risk of fatal or devastating bleeding, the decision to prescribe it should be carefully considered in consultation with critical care specialists when available.

After much debate about its efficacy and questionable FDA approval, activated protein C (trade name Xigris) was recently pulled from the market. Because of the lack of conclusive benefit demonstrated in subsequent studies after the original PROWESS study, the European Medicines Agency felt the risk-benefit ratio of activated protein C needed to be clarified and the maker of Xigris, Eli Lilly, was asked to conduct a new placebo-controlled trial for patients with septic shock. The study was named PROWESS-SHOCK and its results were recently revealed, failing to meet the primary endpoint of statistically significant reduction in 28-day all-cause mortality in patients treated with activated protein C (Xigris). The risk of severe bleeding events was similar in both arms, suggesting no harm. These results call into question the overall risk-benefit balance of activate protein C for the severe sepsis and septic shock patient population, and Eli Lilly has thus decided to withdraw the product from the market worldwide.

The multidisciplinary environment of the ICU often requires inclusion of many different consultants to aid in the care of critically ill patients, as in the setting of sepsis, severe sepsis, and septic shock. A team that includes critical care nursing, intensivists, infectious disease specialists, nephrologists, endocrinologists, surgeons, palliative care experts, nutritionists, social workers, and many others may be necessary to optimize the care of patients with sepsis.

Many patients with sepsis will not require the care of an intensivist; however, consultation should be considered for persistent hemodynamic instability, incipient progression to multiorgan failure, and when uncertainty exists regarding diagnosis or management. Septic patients with multiorgan failure or those requiring advanced airway and/or ventilator management likely deserve intensivist involvement in their care. Consultation with an intensivist should also be considered if invasive monitoring is needed including placement of central venous and arterial catheters, and especially if placement of pulmonary artery catheters is contemplated. In addition, intensivist consultation may assist with decision making to prescribe recombinant APC or corticosteroids in the setting of septic shock.

Not uncommonly, consultation with a nephrologist is required for patients with acute kidney injury who may need renal replacement therapy while critically ill.

The multidisciplinary approach required in the care of complex critically ill patients often requires collaboration and consultation of other clinical specialists, including critical care nursing; respiratory care; physical, occupational, and speech therapy; wound care nursing; social services; pastoral care; palliative care; and others. Clinicians caring for patients in the critical care setting should have a low threshold to involve necessary services in the care of their critically ill patients to optimize care quality and efficiency.

Uncomplicated sepsis that is treated appropriately and aggressively will often have very few untoward outcomes and patients may return to baseline within a very short time frame. Sepsis that progresses to severe sepsis or septic shock more frequently leads to a complicated course for the patient (see Table 138-1). Patients will often require invasive procedures including central venous catheter and arterial catheter placement, as well as possible endotracheal intubation and mechanical ventilation to manage airway, refractory hypoxemia, and/or increased work of breathing. Patients with acute lung injury, also known as noncardiogenic pulmonary edema or acute respiratory distress syndrome (ARDS), may require a prolonged course of mechanical ventilation; their lung function typically improves over time, though not back to baseline. Other patients may suffer kidney injury requiring dialysis therapy. The typically reversible acute tubular necrosis may take days, weeks, or months to resolve, and some patients may require chronic hemodialysis permanently. A relatively common neurologic disorder encountered in sepsis is the development of critical illness polyneuropathy, which will typically manifest in difficulty liberating a patient from mechanical ventilation due to profound weakness. This condition usually resolves, but often can take several months. Psychological stress, often described as posttraumatic stress disorder, is related to the common complication of ICU delirium. Daily awakening in the ICU may minimize delirium.

Patients may transition from a closely monitored setting—either intensive care unit or intermediate care unit—to a medical floor when they are hemodynamically stable (ie, off vasoactive medications) and not in need of frequent CVP and ScvO2 measurement. Any septic patient requiring mechanical ventilation, or even noninvasive mechanical ventilation intermittently, should remain in an ICU setting until their gas exchange or work of breathing problems have resolved and they no longer require supportive devices. Discharge from the hospital may occur quickly if a patient has mild sepsis that is treated and resolves quickly. For a more prolonged hospitalization including mechanical ventilation, a patient may require significant physical rehabilitation within an acute or subacute rehabilitation facility before returning to the home environment.

Institutional sepsis protocols likely will standardize and improve overall care for septic patients admitted to hospitals and ICUs. Since outcomes in severe sepsis and septic shock highly depend on expedient administration of evidence-based interventions, institutions and individual clinicians should consider measuring process outcomes (eg, time to antibiotics in sepsis, use of ScvO2 monitoring) as well as patient outcomes to help best delineate areas of heterogeneous care and help drive consistent evidence-based or guideline-based practice to improve overall outcomes.

Sepsis prevention requires early recognition of disease and prompt treatment to prevent progression to severe sepsis and septic shock. A lower threshold for sepsis evaluation and aggressive therapy needs to be applied to immunosuppressed patients, including patients with HIV, asplenia, neutropenia, and patients receiving immunosuppressive agents. Early removal of urinary catheters, central venous catheters, and liberation from mechanical ventilation to prevent ventilator-associated pneumonia will lower the incidence of sepsis. The cost of caring for septic patients exceeds $22,000 per hospitalization, and much opportunity exists to improve outcomes and reduce hospital lengths of stay and costs for a medical condition that bears high morbidity and mortality (Table 138-6).