Identifying sources of infection and establishing source control is an essential component of the workup and treatment of sepsis. Investigation with history, physical examination, laboratory tests, and imaging can in identifying sources of infection. All organ systems have the potential to develop sources of infection. However, there are inherent difficulties presented by some that require additional diligence, namely, urinalysis, chest radiographs, and intraabdominal infections. Interventions include administration of antibiotics and may require surgical or other specialist intervention. This is highlighted by the Surviving Sepsis Campaign with specific recommendations for time to antibiotics and expeditious time to surgical source control.
Key points
- •
Rapid time to antibiotics is important, but they may not themselves provide adequate source control.
- •
A broad differential is essential in septic patients, because the source of infection may not immediately identifiable.
- •
Typical workup strategies in sepsis such as the chest radiograph and urinalysis can be inherently misleading.
- •
Intraabdominal sources of sepsis are difficult to diagnose, treat, and frequently require rapid surgical intervention.
Introduction
The medical concept of source control as a component of treatment for infection dates back as early as the 17th century bc . The Edwin Smith papyrus of ancient Egypt describes 48 medical cases, one of which details methods at the time to assist with the drainage of a chest wall abscess. In modern medicine, the concept of source control continues to be a cornerstone of sepsis management, albeit now far more complex than the astringents and poultices used by the ancient Egyptians to assist with drainage.
Source control in sepsis first requires identification of an infectious source followed by the subsequent interventions used to control specific sites of infection and to modify factors that promote microbial growth or impair host defenses to infection. In the emergency department (ED), this process can be challenging especially because the source of the infection may not be readily apparent. Whereas a chest radiograph, urinalysis (UA), and blood cultures are ubiquitous in the evaluation of sepsis, sometimes further investigation is required as a surgical process (eg, appendicitis, cholecystitis) may be the underlying culprit pathology.
The Surviving Sepsis Campaign (2012) provides an update to current sepsis treatment guidelines, including multiple topics pertinent to source control and their level of evidence to support these recommendations. These include recommendations regarding obtaining cultures, starting antibiotic therapy, and performing surgical source control.
Introduction
The medical concept of source control as a component of treatment for infection dates back as early as the 17th century bc . The Edwin Smith papyrus of ancient Egypt describes 48 medical cases, one of which details methods at the time to assist with the drainage of a chest wall abscess. In modern medicine, the concept of source control continues to be a cornerstone of sepsis management, albeit now far more complex than the astringents and poultices used by the ancient Egyptians to assist with drainage.
Source control in sepsis first requires identification of an infectious source followed by the subsequent interventions used to control specific sites of infection and to modify factors that promote microbial growth or impair host defenses to infection. In the emergency department (ED), this process can be challenging especially because the source of the infection may not be readily apparent. Whereas a chest radiograph, urinalysis (UA), and blood cultures are ubiquitous in the evaluation of sepsis, sometimes further investigation is required as a surgical process (eg, appendicitis, cholecystitis) may be the underlying culprit pathology.
The Surviving Sepsis Campaign (2012) provides an update to current sepsis treatment guidelines, including multiple topics pertinent to source control and their level of evidence to support these recommendations. These include recommendations regarding obtaining cultures, starting antibiotic therapy, and performing surgical source control.
General workup
History
Although obtaining a history is part and parcel of every patient encounter in the ED, unfortunately it may be limited, misleading, or downright unobtainable owing to the patient’s illness. If the emergency physician is fortunate, the patient or family member may relate complaints such as fevers and chills that point more toward an infectious etiology. However, sepsis is the most complex disease entity emergency physicians will encounter because the patient complaints are often nonspecific and may simply be dyspnea or abdominal pain, which are associated with a great number of possible differential diagnoses. On the extreme end of the spectrum, a patient may be incapable of providing any history owing to an acute change in mental status or an underlying comorbid condition such as dementia.
Because of the challenges described in trying to arrive at a suspicion of sepsis, it essential that the emergency physician try to gather as much information as possible from all available resources. Information can be garnered from emergency medical services personnel, a long-term care facility, family, friends, and perhaps most important, previous patient encounters. A careful review of the patient’s medical record can be invaluable, because it may have previous culture results, recent prior surgeries, significant medications (eg, immunosuppressants), previous and recent antibiotics, or past medical history that changes the evaluation (eg, HIV). Taking the extra time to gather information becomes more important with older patients as they notoriously will have nonspecific presentations of sepsis and may be more likely to harbor surgical disease.
Physical Examination
Some patient complaints such as chest pain allow the emergency physician to direct their physical examination in pursuit of specific disease entities on the differential diagnosis. However, it is not uncommon that a patient may present with vague complaints and the physical examination must be appropriately thorough as no one organ system rises to the top as the potential origin of disease. Sepsis often falls into this category and the astute emergency physician should realize that the physical examination is of tremendous importance in guiding diagnostic inquiry.
The clinical manifestations of inflammation suggestive of infection may or not be apparent on physical examination. These include the rubor (redness), calor (warmth), dolor (pain), and functio laesa (loss of function). A thorough examination should be undertaken and any abnormalities noted. General observations such as alertness, distress, and pain are all useful. The examination should take note of any signs of hardware or prior and/or recent surgeries. These could include indwelling venous catheters, port sites, drains, urinary catheters, or any other medical devices. Also consider scars that could be suggestive of underlying hardware, including orthopedic devices.
Although the emergency physician is likely accustomed to considering pulmonary and urinary sources of sepsis, it is important to keep in mind that the skin (eg, cellulitis, abscess), genitourinary (eg, abscess), joints (eg, septic joint), and abdomen (ie, numerous disease processes) may all yield information that suggests further inquiry is required. It is also important to mention that if the initial physical examination is not contributory, consider repeat examinations, especially of the abdomen, because the findings may change over the course of minutes to hours in the ED.
Laboratory Tests
Serum laboratory studies will most certainly be an integral component in the evaluation of sepsis, yet they are fraught with poor sensitivity and a potentially significant time delay. Reasonable laboratory tests to send initially include a complete blood count with differential, a complete metabolic profile, which will include liver functions tests, a UA and a point-of-care lactate (see Brit Long and Alex Koyfman’s article, “Ready for Prime Time? Biomarkers in Sepsis,” in this issue). Other studies will depend on the patient, current medications (ie, coagulation studies for patient on warfarin), comorbid conditions (ie, B-type natriuretic peptide for patient with congestive heart failure history), and the clinical question the emergency physician desires to explore.
Cultures
Cultures are a vital part of the inpatient management of patients with sepsis. Unfortunately, there remains a large percentage of patients who will have “culture-negative” sepsis. Literature shows rates of negative cultures for septic patients ranging from 40% to 60%; pulmonary infections have the highest rate of culture-negative results. The inpatient team is then left to decide which antibiotic to discontinue and when to narrow the antibiotic spectrum without the benefit of a positive culture. This highlights the importance of optimizing the potential to obtain accurate and contributory cultures. Blood cultures should ideally be obtained before the start of antibiotic therapy, as long as this can be done in a timely fashion (<45 minutes). This practice is a bit of a contradiction in sepsis management in that the guidelines and recent evidence strongly suggest that early, broad-spectrum antibiotic therapy is tremendously important for improved patient outcome. However, cultures are essential for the inpatient team to narrow antibiotic therapy and avoid associated consequences of prolonged broad-spectrum antimicrobial therapy (eg, antimicrobial resistance and Clostridium difficile ). It is incumbent on to the emergency physician to communicate with his patient care team (eg, nurses and techs) about the dual priorities of obtaining blood cultures while not excessively delaying antimicrobial therapy. The emergency physician may have to consider alternative methods for obtaining cultures, such as a radial artery blood draw or perhaps a femoral straight stick to expedite collection.
Blood cultures should always include 2 sets so that both aerobic and anaerobic bottles are obtained. In any patient with an indwelling central venous catheters, one of the culture sets should be drawn through this line to help potential identification of infected devices. Urine cultures should be sent routinely for any patient suspected to have sepsis because interpretation of the UA is notoriously unreliable. If the patient is producing sputum and a pulmonary source is suspected, efforts should be made to induce sputum for gram stain and culture. Consideration should be given to wound cultures as appropriate especially if purulent drainage is evident on examination or able to be expressed. Cerebrospinal fluid should be obtained as appropriate and once again antibiotics should not be significantly delayed to obtain cerebrospinal fluid. Specifically with regard to blood cultures, ideally at least 10 mL of blood should be cultured per tube; this volume has been shown to substantially increase detection rates of bloodstream infections (BSIs). Volumes of blood should also be equal between culture tubes, which may assist with the identification of indwelling devices as a source. For instance, if cultures drawn through an indwelling central line are positive more than 2 hours before those from a fresh peripheral blood draw, this finding suggests that the indwelling device is the source of the infection.
Finally, if the patient remains in the ED for a number of hours and has an abrupt onset of fever with or without rigors, this is an ideal time to repeat blood cultures as the patient may be bacteremic. Additional blood cultures will increase the yield of culture results, but also increase the rate of false-positive results, so it is important to consider this when ordering repeat cultures without a change in patient clinical status.
Antibiotics
Antibiotic administration is covered in detail in Michael G. Allison, Emily L. Heil and Bryan D. Hayes’s article, “Appropriate Antibiotic Therapy,” in this issue. Although narrow spectrum antibiotic therapy is the goal in the inpatient setting when the patient is clinically stable and culture results can be used to guide therapy, in the ED it is “empiric” therapy that guides antimicrobial decision making. The emergency physician needs to use their hospital’s antibiogram to determine what is the most appropriate broad-spectrum agent of choice in the initial management of severe sepsis and septic shock ( Fig. 1 ). Although this article focuses on source control, it is rare that definitive patient care is achieved without a combined approach between source control and antbimicrobial therapy when a source of sepsis is able to be targeted and amenable to surgical or other invasive management.
Imaging
The ED evaluation of sepsis is very likely to involve imaging, although it is important to keep in mind that the use of imaging should be directed at specific potential sources when clinical suspicion warrants the resource utilization, cost, and potential radiation. Although a chest radiograph is essentially ubiquitous in the sepsis evaluation, ultrasound (US) imaging, computed tomography (CT) scanning, and MRI are less commonly used and associated with more cost, resource allocation, and potential radiation exposure (CT). Perhaps the most important concept to consider when ordering advanced imaging is that the patient will be out of the ED and other important facets of sepsis evaluation and management will be interrupted. For example, if the emergency physician orders a head CT early in the evaluation of a patient with altered mental status potentially owing to sepsis, the patient may quickly be taken out of the ED for this and their antibiotics, blood cultures, and intravenous fluids administration will be delayed. On the other side of the spectrum, it is always good medicine to consider other sources of sepsis beyond the routine sources evaluated by chest radiograph, UA, and blood cultures. The astute emergency physician will pay close attention to the abdominal examination to evaluate for cholecystitis, appendicitis, or noninfectious intraabdominal pathology that may mimic sepsis (eg, hollow viscous perforation, small bowel obstruction). A quick bedside US examination is noninvasive and an easy method to evaluate for biliary pathology. Repeat examinations can be invaluable in guiding the emergency physician toward a decision on advanced imaging, especially if the patient is sick and the chest radiograph and UA do not suggest a clear source of infection. Older patients are especially likely to have more atypical presentations and earlier use of advanced imaging should be considered especially if intraabdominal pathology is suspected.
Surgical Source Control
There are 3 cardinal principles when it comes to obtaining surgical source control. These include drainage of infected fluid collections, debridement of infected tissues and/or removal of related foreign bodies, and measures to definitively correct anatomic dysfunctions resulting in ongoing infectious contamination.
Certain specific types of infection may require more urgent mechanical source control than others. These include necrotizing soft tissue infections, peritonitis, and ischemic bowel. Interventions should be undertaken to achieve source control within 12 hours of diagnosing sepsis related to a source that will require an invasive corrective measure. Other specific recommendations regarding source control include using the least invasive procedure possible to achieve source control as well as removal of any vascular devices that are suspected as possible sources of infection. The caveat regarding vascular devices is that if the patient is unstable, the indwelling vascular device may be necessary for resuscitation and this takes priority over source control until hemodynamic stability is achieved.
Further Investigation
Rapid identification in the ED of a specific source of infection can be difficult. Studies have suggested that correct source identification in the ED, even with full diagnostic testing, was accurate in only 85% of pulmonary, 80% of urogenital, and only 68% on of intraabdominal sources. This further reinforces the concept of initiating broad spectrum antibiotics in the ED and allowing the inpatient team to narrow the antibiotic regimen with the benefit of culture results and further investigative workup. Owing to some of the inherent misinformation that can be presented by basic workup of UA and chest radiographs, as discussed, the provider should always keep an open mind regarding the possibility of other sources.
Sources
Central Nervous System
The central nervous system (CNS) is a concerning, albeit infrequent source of sepsis. One study found as few as 1% of presenting septic patients had a CNS etiology. Possible sources include meningitis, encephalitis, brain abscess, and epidural abscesses. Although relatively rare, they have high rates of mortality and morbidity. Adult bacterial meningitis has a mortality rate of approximately 25%, and 20% to 30% of survivors will have neurologic sequelae. Typical presenting features include headache, fever, nuchal rigidity, and altered mental status. Although it is unusual for patients to present with all 4 of these classic findings, nearly 95% of patients will present with at least 2 of them.
All patients with clinical suspicion for a CNS source of infection should have a lumbar puncture (LP) performed unless it is contraindicated. A head CT scan before any LP is recommended (by the Infectious Diseases Society of America) for patients who are immunocompromised (eg, AIDS), have a history of prior CNS disease, new-onset seizures, papilledema, abnormal level of consciousness, or a focal neurologic deficit. They also specifically recommend giving empiric antibiotics after obtaining blood cultures, but before proceeding with the LP. That is, do not delay administration of antibiotics at all to obtain the LP. Dexamethasone (10 mg) should be administered intravenously 15 to 20 minutes before giving the antibiotics because it has been shown to reduce morbidity and mortality of bacterial meningitis.
Pulmonary
Respiratory complaints are common presentations in numerous disease entities. In terms of infectious etiologies, these complaints can result from mild upper respiratory infections to severe sepsis or septic shock secondary to pneumonia. The most common presenting symptoms of pneumonia include shortness of breath, chest pain, and cough. One study found at least 96% of patients with pneumonia presented with at least 1 of these complaints.
The chest radiograph is the standard imaging modality used for the diagnosis of pneumonia. Nearly 19% of all visits to an ED include a chest radiograph. It should be noted, however, that these studies can frequently be misleading. Using CT imaging as a standard for comparison, 27% to 43% of patients with an infiltrate on CT imaging will have a negative or nondiagnostic chest radiograph. Possible reasons for this include poorly visualized regions on the chest radiograph (eg, lung bases, lingual), very small infiltrates, or early infection that has not yet led to radiographic change. Conversely, a suspected infiltrate on chest radiograph may actually represent pulmonary edema, chronic lung disease, or delay in radiographic resolution of prior infection. Again, using CT imaging as a standard for comparison, one study found 27% of patients with an infiltrate on chest radiography did not have concordant findings on CT scanning. The reasons for false-positive results are also varied and include atelectasis, poor inspiration, or poor positioning.
Abdomen
Abdominal pain is a common chief complaint in the ED and accounts for 5% to 10% visits. Intraabdominal infections pose a unique challenge to the emergency physician because they vary in presentation and severity of infection, are frequently polymicrobial, and nearly always require some type of surgical intervention. It is therefore not surprising that mortality rates from intraabdominal infections vary widely from less than 1% to greater than 90%. Intraabdominal infections can be divided into 2 major classes, uncomplicated and complicated. Uncomplicated intraabdominal infections are those that are confined to a single organ with no anatomic disruption. Complicated intraabdominal infections occur when an infectious process extends beyond a single organ. These lead to abscesses, peritonitis, and anatomic disruption.
The diagnostic accuracy of a combined history, physical, and laboratory studies in the evaluation of acute abdominal pain is a paltry 43% to 59%. This level of accuracy does not depend on the level of training (ie, resident or attending or even specialist). Thus, it is not unreasonable to have a low threshold for advanced imaging in patients in whom the emergency physician is concerned for an intraabdominal source of sepsis. Conventional radiology (eg, plain films) is of little benefit beyond cases of bowel obstruction. US imaging is a great first option if biliary pathology is suspected. If the emergency physician is proficient, this can be accomplished at bedside. It also avoids the radiation and contrast exposure of CT imaging. CT imaging, however, has the greatest sensitivity and specificity of all imaging modalities, correlating with the correct final diagnosis in 62% to 92% of cases. When ordering CT imaging to evaluate for intraabdominal pathology, it is typically not necessary to use oral contrast. Multiple studies have shown that intravenous contrast is adequate for diagnostic accuracy in abdominal CT imaging and dramatically increases turn around time. It is worth mentioning that patients with severe sepsis or septic shock are often hypovolemic and may have acute kidney injury so any consideration of intravenous contrast should be weighed carefully.
In general, intraabdominal infections benefit from rapid source control; delays in definitive management are closely related to increases in morbidity and mortality. Certain intraabdominal infections require rapid surgical source control; these infections include but are not limited to necrotizing soft tissue infections, peritonitis, appendicitis, cholecystitis, and ischemic bowel. However, others, specifically pancreatic necrosis, often benefit from delayed surgical intervention. Early surgical consultation is of tremendous importance in cases of sepsis owing to an intraabdominal source.
Unfortunately, it is not uncommon to miss an intraabdominal source of sepsis, especially in the elderly. The elderly and those with altered mental status may not always have a history or examination that suggests an intraabdominal infection. These patients may wind up being treated on a medical service and their definitive surgical intervention is significantly delayed, resulting in significant morbidity and mortality. The astute emergency physician will use repeat examinations and appropriate advanced imaging when a source is not readily identified (eg, pneumonia) on the initial patient examination.
Genitourinary
Urinary tract infections (UTIs) are a frequently encountered source of sepsis, also known as urosepsis. The percentage of sepsis presentations attributed to the urinary tract varies widely in the literature, ranging from 8% to 59%, with an average compiled from 16 studies (8667 patients) of 13%. Much of this variation has been attributed to the age range of the patients studied. It is a more common source in older patients and is the most common source in patients greater than 65 years of age. The greatest single risk factor is an indwelling urinary catheter, but other risk factors include incontinence or neurologic disorders.
Historical factors associated with an increased likelihood of a UTI include fevers, rigors, dysuria, frequency, retention, and physical examination findings of suprapubic or flank tenderness. Other historical factors to consider include frequent prior UTIs, recent urinary tract instrumentation, increased confusion, and falls. All patients presenting with a concern for sepsis should have a UA performed. A urine culture should be sent on all patients admitted for severe sepsis and septic shock.
Urine culture is considered the gold standard for diagnosis of UTI; however, this requires 24 to 48 hours to yield results. The emergency physician must rely on the urine dipstick and microanalysis findings when considering an infection. Typically these are nitrite, leukocyte esterase, white blood cells, and bacteria. The sensitivity and specificity of these tests vary widely in the literature and thus they can often be misleading. See Table 1 for more information.
