Sepsis is an important diagnosis in the critical care unit. This chapter will discuss its definition, pathogenesis, early goal directed therapy, and types of infection.

Systemic inflammatory responses (SIRS) is defined as more than 2 of the following: temperature less than 36°C or greater than 38°C, heart rate (HR) greater than 90 beats/min, respiratory rate (RR) greater than 20 breaths/min, PaCO₂ less than 32 mmHg, white blood cell (WBC) count less than 4000 per mm³ or more than 12,000 per mm³, or greater than 10% immature bands.¹ Systemic inflammatory responses associated with infection have been the definition of sepsis for 2 decades. Sepsis with organ dysfunction has been the definition of severe sepsis. Septic shock has been the definition of hypotension despite adequate fluid resuscitation requiring vasopressors and attributed to sepsis.

Improved understanding of this syndrome and the need for uniformity prompted a reevaluation of the criteria. From the Third International Surviving Sepsis Guidelines, the definition of sepsis is defined as life-threatening organ dysfunction by a dysregulated host response to infection.^2,3 Organ dysfunction is defined as an acute change of 2 points or more in total Sequential Organ Failure Assessment (SOFA) score secondary to infection.^2,3 Mortality is 10%. A bedside SOFA or quick SOFA (qSOFA) consists of RR of 22 or more per minute, altered mental status, and systolic blood pressure (SBP) of 100 mmHg or less. The full SOFA score is in Table 18-1.⁴

Septic shock is defined as persistent hypotension and lactic acid of 2 mmol/L (18 mg/dL) or greater despite adequate fluid resuscitation, and the need for vasopressors to maintain mean arterial pressure (MAP) of greater than 65 mmHg.^2,3,5 Mortality is 40%.

The pathogenesis of sepsis is complex, and this is a very brief summary. It is a response of the host to an infection with proinflammatory and anti-inflammatory components.⁶ The proinflammatory response attempts to kill the infective agent but consequently causes damage to the host tissue. The anti-inflammatory response attempts to mitigate the proinflammatory response but predisposes the host to secondary infections.

The paradoxical proinflammatory and inflammatory response of sepsis begins with identification of the infective agent through innate immunity. This is achieved via 4 groups of receptors (toll-like receptors, c-type lectin receptors, retinoic acid inducible gene 1–like receptors, and nucleotide-binding oligomerization domain–like receptors) that recognize pathogen-associated molecular patterns in infectious agents and damage-associated molecular patterns in injured cells.⁷ There is an altered coagulation that leads to disseminated intravascular coagulation.⁸ Organ dysfunction is partially secondary to impaired tissue oxygenation from hypotension, reduced red-cell deformability, and microvascular thrombosis. Humeral, cellular, and neurogenic mechanisms diminish the inflammatory response. This often leads to apoptosis of B cells, CD4 T cells, and follicular dendritic cells, which leads to further immunosuppression.^9,10

Early goal-directed therapy (EGDT) has been the gold standard for the management of sepsis.¹¹ However, current studies have challenged the benefits of EGDT. The Protocolized Management in Sepsis (PROMISE) trial showed that 184 out of 623 patients in EGDT and 181 out of 620 patients in the usual care group had died (relative risk in EGDT, 1.01; 95% confidence interval [CI], 0.85–1.20; P = 0.90) for absolute risk reduction (ARR) of −0.3 percentage points (95% CI, −5.4 to 4.7) and no significant differences in quality of life or rates of serious adverse effects.¹² Early goal-directed therapy increased costs, with cost-effectivity less than 20%.¹² The Australasian Resuscitation in Sepsis Evaluation (ARISE) and Australian and New Zealand Intensive Care Society (ANZICS) group showed that in 90 days after randomization, EGDT had 147 deaths and the usual care group had 150 deaths, with rates of 18.6% and 18.8%, respectively (absolute risk difference, −0.3 percentage points; 95% CI, 0.4 to 3.6; P = 0.90) and no difference in survival time, hospital mortality, duration of organ support, and length in hospital stay.¹³ However, EGDT is still advocated, since there is no harm in the interventions, and clinicians need guidance in managing sepsis.

Resuscitation

Fluid resuscitation (30 mL/kg^11,12 or 2 Liters¹³) is recommended within the first 3 hours. Certain patients may require more fluid as determined by dynamic or functional hemodynamic monitoring, such as stroke volume variation (SVV), over static monitoring, such as central venous pressure (CVP). Central venous pressure is limited due to its inability to predict fluid responsiveness in patients with normal values (8–12 cm H₂O).^14,15 Crystalloids are preferred over colloids since crystalloids are cheaper and colloids have no clear benefit² (Strong Recommendation and Moderate Quality of Evidence). Well-balanced crystalloids are advocated, since hyperchloremia is thought to cause decreased glomerular filtration rate (GFR) due to constriction of afferent arteriole.¹⁶ However, the 0.9% saline vs plasma-lyte 148 for icu fluid therapy (SPLIT) trial suggested that use of buffered crystalloid over normal saline did not reduce risk of acute kidney injury (AKI). Renal replacement therapy (RRT) in the buffered crystalloid group is 38 of 1152 [3.3%] compared to 38 of 1110 [3.4%] in the saline group with absolute difference −0.1%; 95% CI, 01.6% to 1.4%).¹⁷ However, a recent cluster-randomized, multiple-crossover trial in 5 intensive care units (ICUs) which compared balanced crystalloids such as lactated Ringer’s solution or Plasma-lyte A versus normal saline showed 14.3% of balanced crystalloid group had major adverse kidney event compared to 15.4% in saline group (marginal odd’s ratio 0.91; 95% confidence interval. 0.84–0.99; conditional odd’s ratio 0.90; 95% CI 0.82–0.99; P=0.04).¹⁸ In hospital mortality at 30 days was also less in the balanced crystalloid group (10.3%) compared to saline group (11.1%); incidence of new RRT was less in balanced crystalloid group (2.5%) compared to saline group (2.6%)(P=0.06).¹⁸

The use of albumin over crystalloid does not show consistent mortality benefit in multiple studies.^2,19-24 Hetastarches are associated with higher rates of RRT and their mortality benefit is variable^2,25 (Strong Recommendation and High Quality of Evidence). Gelatin, another colloid, did not increase mortality or AKI compared to albumin or crystalloid.²⁶

The goal of fluid resuscitation is to maintain MAP above 65 mmHg and normalize lactate.² Mean arterial pressure is the driving force for tissue perfusion and at an MAP less than 65 mmHg, tissue perfusion becomes linearly dependent on arterial pressure.² Higher MAP goals have not been associated with improved mortality but have been associated with a higher risk of arrthymias.² Lactate is an indirect measure of perfusion, and early lactate clearance strategy has been shown to reduce mortality.^27-30

Vasopressors can be utilized to meet these goals if fluid resuscitation is insufficient. Norepinephrine is the preferred first line vasopressor (Strong Recommendation and Moderate Quality of Evidence). Dopamine can be used in lieu of norepinephrine in patients who are bradycardic with a low incidence of tachyarrthymias. If insufficient, vasopressin or epinephrine can be added (Weak Recommendation and Low Quality of Evidence). Dobutamine can be added if the patient has a persistent hypoperfused stated despite adequate fluid resuscitation and vasopressors (Weak Recommendation and Low Quality of Evidence).

Norepinephrine is preferred to dopamine because it has a more potent effect in increasing cardiac output, it is less protachyarrythmic, and it has no known association with immunosuppression.³¹ A metanalysis of 11 randomized trials showed that norepinephrine had lower mortality (risk ratio, 0.89; 95% CI, 0.81–0.98) and lower arrhythmia risk (risk ratio, 0.48; 95% CI, 0.40–0.58).³² It has been suggested that there is no difference in mortality with the use of epinephrine versus norepinephrine, but epinephrine may increase lactate production via β₂-adrenergic receptors in the skeletal muscle.² It also has been suggested that there is a relative vasopressin deficiency in septic shock, but most of the data regards vasopressin as a sparing effect to norepinephrine.² Dobutamine is considered a first-line inotropic agent to improve oxygen delivery. It has been shown to improve clinical outcomes, lactate, and central venous oxygen saturation (SCVO₂).² Phenylephrine causes splanchnic vasoconstriction and its effects on clinical outcomes have yet to be determined.³³ Low-dose dopamine does not improve urine output, RRT, intensive care unit stay, hospital stay, arrhythmias, and survival compared to placebo, and is generally not advocated.²

Antibiotics

Intravenous (IV) antibiotics should be administered within 1 hour of recognition of sepsis because each hour delay increases mortality, acute kidney injury, acute lung injury, and length of stay^34-37 (Strong Recommendation and Moderate Quality of Evidence). However, a recent meta-analysis suggested that there is no mortality benefit to administering antibiotics within 3 hours versus within 1 hour of sepsis or septic shock recognition.³⁸ Considerations for the choice of antibiotics include anatomic site of infection, prevalent pathogens in the community and resistance patterns, and presence of immunodeficiencies.² Duration of antibiotics is typically 7 to 10 days but can be longer if there is a slow response, undrainable foci, Staphylococcus aureus bacteremia, a fungal infection, a viral infection, or states of immunodeficiency² (Weak Recommendation and Low Quality of Evidence). Daily assessment of antibiotic de-escalation should take place with clinical evaluation and culture results with or without utilization of biomarkers such as procalcitonin (PCT).

The patient should have cultures if there is no substantial delay to initiation of antibiotics (45 minutes).² Culture-negative sepsis is not uncommon. A prospective observation cohort suggested that culture-negative sepsis was associated with women, fewer comorbidities, lower Acute Physiology and Chronic Health Evaluation (APACHE) score, lower SOFA score, lower procalcitonin levels, less tachycardia, higher blood pressures, shorter duration of stay, and lower ICU mortality compared to culture-positive sepsis.³⁹

Corticosteroids

Hydrocortisone 200 mg IV daily is suggested for septic shock despite adequate fluid resuscitation and vasopressor support² (Weak Recommendation and Low Quality of Evidence). The recommendation is weak due to the inconsistencies regarding whether it improves mortality.^40-47 Random cortisol levels are not useful, since they may underestimate or overestimate true values.²

Community-acquired pneumonia (CAP), nervous system infections such as meningitis, encephalitis, abscess, neuritis, infective endocarditis (IE), pericardial infections, skin and soft tissue infections, and infections during pregnancy are discussed in this chapter. Other types of infections are discussed in the following chapters: Chapter 19, “Healthcare-Acquired Infections”; Chapter 20, “The Immune System and Infection”; and Chapter 21, “Antimicrobials.”

Community-Acquired Pneumonia

Clinical presentation of pneumonia can consist of fever, shortness of breath, cough, pleuritic chest pain, productive purulent or clear sputum, nausea, vomiting, diarrhea, or altered mental status, but the gold standard for confirmation is chest radiograph or chest computed tomography (CT). Typical pathogens associated with community acquired pneumonia are Streptococcus pneumoniae, Haemophilus influenzae, S aureus, group A streptococci, Moraxella catarrhalis, aerobic Gram-negative bacteria, atypical bacteria (Mycoplasma pneumoniae, Chlamydia pneumoniae, C psittaci, and Legionella), and viruses. In outpatient management of CAP, cultures are optional due to difficulty in obtaining good specimens, costs, and concerns for delay of treatment.⁴⁸ Cultures including sputum, blood, polymerase chain reaction (PCR), and urinary antigens for S pneumoniae and Legionella are advocated in hospitalized patients, especially for patients with severe pneumonia.

Site-of-care decision is based on risk of mortality, and from the observation that patients with less-severe CAP who are treated as outpatients generally recover faster, with less cost and less exposure to more resistant pathogens. Pneumonia severity index (PSI), CURB-65 (confusion, urea level, respiratory rate, blood pressure, and age less than 65 years), SMART-COP (systolic blood pressure, multilobar involvement, albumin level, respiratory rate, tachycardia, confusion, oxygenation, and arterial pH), and Severe Community-Acquired Pneumonia (SCAP) score are used to stratify risk and guide treatment⁴⁸ (Tables 18-2 and 18-3).

Nervous System Infections

In addition to neuroimaging, cerebrospinal fluid (CSF) analysis is crucial to identifying infectious and noninfectious pathologies of the CNS. There are 4 entities with normal CSF analysis: (1) early bacterial meningitis, (2) cryptococcal meningitis with HIV, (3) parimeningeal foci, and 4) herpes simplex encephalitis. Most common method to obtain CSF sample would be through a lumbar puncture. Prior to performing this procedure, a computed tomography of the head must be obtained to rule out any space occupying lesion. Presence of seizures could also indicate transient increase in intracranial pressure; therefore, the lumbar puncture should be delayed by 30 minutes or more. Its presence could suggest the risk of herniation if the lumbar puncture is performed.

Cerebrospinal fluid is produced in the choroid plexus in the lateral, third, and fourth ventricles, circulated in the subarachnoid space, and reabsorbed in the arachnoid villi. The normal volume is 150 mL and produced at 20 mL/h. The normal CSF glucose–to–blood glucose ratio is 0.6 and glucose enters the CSF through choroid plexus and capillaries. The CSF glucose is decreased by consumption of white blood cells and organisms and/or decrease transfer from inflamed meninges. Protein generally is excluded in CSF but levels increase if there is a disruption of the blood–brain barrier. The normal levels along with the levels seen in several conditions are seen in Table 18-4.^55-57

Encephalitis

Encephalitis is an inflammatory condition of the brain parenchyma secondary to an infectious or noninfectious cause. In infections, neural cells of gray matter are the target with associated perivascular inflammation, neuronal destruction, and tissue necrosis. It might be difficult to distinguish meningitis and encephalitis, and it is not uncommon that both processes can occur together and can have associated CSF pleocytosis called meningoencephalitis. In autoimmune or postinfectious causes, there is perivenular inflammation and demyelination of the white matter. Generally, patients with encephalitis will have abnormal brain function such as altered mental status, neurologic deficits, and personality changes, and may have abnormal radiographic findings.

Noninfectious causes include acute disseminated encephalomyelitis (ADEM), which is encephalitis triggered by an autoimmune response to prior antigen exposure. Prior antigen exposure can include immunization (1–14 days after vaccination) or febrile illness up to 1 week after the onset of a rash.⁵⁸ Although bacterial, protozoal, and parasitic infections can cause encephalitis, the most common infectious causes in the United States include herpes simplex, West Nile virus, enteroviruses, and herpesviruses.⁵⁸

Investigation begins with an epidemiologic review of the patient’s history, extraneurologic manifestations such as rash, and laboratories such as cultures, polymerase chain reaction, CSF analysis, and magnetic resonance imaging (MRI). Certain MRI patterns can suggest etiology. Herpes simplex encephalitis has temporal lobe involvement; flaviviruses and the Eastern equine encephalitis virus have mixed or hypodense lesions on T1 images of the thalamus, basal ganglia, and midbrain, and hyperintense lesions on T2 images; enterovirus 71 encephalitis has hyperintense lesions on T2 imaging and flair lesions in the midbrain, pons, and medulla; and ADEM has multifocal abnormalities usually in the subcortical white and sometimes even gray matter on T2 and fluid attenuation inversion recovery (FLAIR) sequences⁵⁸(Table 18-5).^59-93

Meningitis

Bacterial meningitis work-up should begin with CSF analysis, including a Gram stain, which identifies 60% to 90% of bacteria with a specificity of greater than 97%.^94,95 The likelihood that Gram stain can identify the bacteria is dependent on the concentration and the specific pathogen: S pneumoniae > H influenza > Neisseria meningitidis > Gram-negative bacilli > Listeria.^96-98 Guidelines do not recommend routine use of latex agglutination tests although it can be used if the Gram stain is negative. Agglutination may be most useful for patients who have received empiric antibiotics and have negative Gram stain and culture. The limulus amoebocyte lysate assay has been suggested as a diagnostic tool to identify the presence of Gram negative infections; however, the assay rarely influences clinical practice and routine use is not suggested. Polymerase chain reaction testing may be useful in Gram stain and culture-negative cases in which suspicion for bacterial meningitis remains high.⁹⁴

Meningitis is a neurologic emergency, and antibiotics should be started as soon as possible. Antibiotic regimen and duration are shown in Tables 18-6 and 18-7.^94-112 Adjunct dexamethasone therapy (0.15 mg/kg every 6 hours for 2–4 days with the dose administered 10–20 minutes before or concurrent with the first dose of the antimicrobial) in adults with suspected or proven pneumococcal meningitis is based on a study that showed unfavorable outcome and death was less in the dexamethasone group (26% vs 52%, P = 0.006 and 14% vs 34%, P = 0.02).¹⁰¹ Adjunct dexamethasone may prevent hearing loss in patients with H influenzae. Ampicillin should be added to cover Listeria if there is advanced age, alcoholism, or immunocompromised states.