Systemic Infections in the Neurologic Intensive Care Unit
Evidence-based management guidelines have been reported in recent years, with the hope of inspiring effective and costefficient algorithms for the work-up and treatment of infectious diseases in the critical care unit.1–7 Thoughtful application of these guidelines in the neurologic intensive care unit (ICU) is appropriate and important, given the impact of infection and fever on patient outcomes. We intend to present concise, guideline-based management principles for the critical care unit, while addressing idiosyncrasies in the care of neurologically injured patients.
General Principles of Fever Management
Body Temperature and Neurologic Outcome
Hyperthermia and hypothermia are both important clinical indicators that an infectious process may be present in a critically ill patient. In units caring for neurologically injured patients, body temperature has a clear association with neurologic outcome.8 Hyperthermia within the first 24 hours has a negative impact on neurologic outcome in stroke, probably because of increased metabolic demand in the ischemic pneumbra.9 Conversely, hypothermia is associated with increased morbidity, particularly in the context of infection, hypotension, and arrhythmias.10 Ultimately, neurologic patients probably benefit from induced normothermia, which prevents the hypermetabolic state of hyperthermia while avoiding the complications associated with hypothermia. Avoiding hypothermia requires the close monitoring of patient temperature and the establishment of appropriate triggers to prompt an infectious work-up, appropriate systems-based identification of infectious and noninfectious sources, and measures to prevent hypothermia or hyperthermia.
Monitoring
In order of preference, temperature monitoring should be achieved by intravascular, esophageal, or bladder thermometry. If these methods are not available, measurements can be collected at rectal, oral, or tympanic membrane sites. Axillary measurements, temporal artery estimates, and chemical dot thermometers should not be used.3
Triggers Prompting Infectious Work-up
Although temperature is an important indicator of possible infectious status, automatic laboratory or radiologic investigations based on temperature alone should be avoided. The new onset of a temperature of 38.3°C or higher or of 36.0°C or lower in the absence of a known cause of pyrexia or hypothermia is an appropriate trigger for clinical assessment.3 Careful clinical assessment drives decision making. Routine “panculturing” should not be performed without a clinical assessment, and microbiological specimens should be obtained at various time points for culture. The likelihood of infectious versus noninfectious processes arises from the clinical circumstance, and the site of infection may become apparent from the history and physical examination. Antibiotics should not be considered the antipyretics of choice3 ( Table 19.1 ).
Suspected Bloodstream Infection
Bloodstream infections are a serious problem, and any suspicion of this type of infection should be investigated thoroughly and adequately addressed. It is very important that the clinician and laboratory personnel communicate effectively and regularly regarding infection evaluation protocols to improve their diagnostic value and to reduce technical errors as technology rapidly evolves.12 Some important clinical triggers for obtaining a blood culture are listed in Table 19.2 . Standing orders should not be written for obtaining blood cultures; they should be indicated either after clinical assessment or in the presence of a known bloodstream infection. Such test-of-cure cultures, when indicated, may be discontinued after three consecutive days of negative cultures.6
Blood cultures are most useful if they are drawn during a febrile episode, especially before the fever dissipates. However, in the case of suspected endocarditis or the more common intravascular device–related sepsis, blood cultures should be drawn as soon as possible. When attempts have been made to achieve normothermia in a patient, blood cultures should be considered for an elevation in the patient′s white blood cell count. These blood cultures ideally are drawn from two separate sites by percutaneous venipuncture. Samples both by venipuncture and from an existing vascular catheter are preferred only when vascular access has exceeded 4 days. Inflammation at the catheter exit site is not a reliable predictor of intravascular catheter–related infection because the contamination of catheter lumens is often involved in the pathogenesis of intravascular infections.13 If venous access is not an option, samples should be drawn from two separate vascular catheters and carefully labeled.6
Fever and any one of the following: |
|
|
|
|
|
|
|
|
|
Test of cure: |
|
|
|
|
Blood cultures for bacteria and yeast should consist of two samples in aerobic and anaerobic bottles from each site. Current blood culture media allow the detection of yeast without the need for special fungal blood cultures. Quantitative blood cultures in which special isolator blood culture tubes are used may be indicated for the diagnosis of bacteremia associated with cuffed Hickman or Broviac catheters or subcutaneous central venous ports.6 The more common method of diagnosis of catheter-related infection is the differential time to positivity of blood cultures taken from catheter and peripheral blood sites. When the differential time exceeds 2 hours between the catheter culture and the peripheral culture, the inside of the catheter is usually considered the source of bacteremia, and the catheter should be removed if possible.3
Table 19.3 summarizes some important guidelines to be considered when blood cultures are obtained. Additional recommendations are to avoid collecting from more than two sites at a time, collecting more than three cultures per day, and collecting more than four cultures per febrile episode.
Diagnosing and treating the cause of a fever in a patient with catheters can be complicated, but a standardized management algorithm has been proposed in Fig. 19.1 .
Treat resulting cultures per the Infectious Diseases Society of America guidelines reported by Mermel et al for the catheters and pathogenic species involved.6 Although the tips of long intravascular catheters, such as percutaneously inserted central venous catheters, are often sent for culture, the preferred segment of the catheter to send to the laboratory is the “intracutaneous” segment. This is the catheter segment just inside the blood vessel that is not contaminated by the skin. The pathophysiology of external (rather than intraluminal) colonization and infection has a bacterial gradient highest at the intracutaneous site and lowest at the catheter tip. Hence, more falsenegative results are likely to be obtained by sending the long line catheter tip than by sending the intracutaneous segment.
Suspected Noninfectious Fever
Noninfectious fevers can often be difficult to diagnose, necessitating a careful and thorough history and physical examination. All recent medication changes and blood products received should be considered in the febrile patient. Establishing a temporal relationship between the initiation or change of a suspected drug and the fever can be helpful in establishing a causal relationship. If a particular medication is suspected as the causative agent, it should be stopped, or a comparable substitution should be made. Drug-induced fevers may require several days to resolve, depending on the drug pharmacokinetics and patient profile. Phenytoin is notorious for causing drug fever and hepatitis in neurosurgical patients. Some nonpharmacologic, noninfectious causes of fevers are listed in the last column of Table 19.1 .
Suspected Postoperative Fever
A chest radiograph, urinalysis, and urine culture are usually not necessary for fever during the first 72 postoperative hours if the fever is observed as an isolated event.3 Patients with an indwelling bladder catheter and fever for longer than 72 hours postoperatively should have a urinalysis, and a blood culture should be drawn.3 The careful daily examination of surgical wounds is important in the setting of fever, and cultures should be obtained if signs of symptoms of infection exist. In any case of postoperative fever, clinicians should be suspicious of deep vein thrombosis, superficial thrombophlebitis, and pulmonary embolism.3 Suspicion for a thrombotic fever should be greater in sedentary patients, those with malignancies, those on oral contraceptives, or those with limb immobility.3
Fever assessment in the neurologic ICU must include a daily examination of the surgical site for signs of infection, such as erythema, purulence, swelling, and tenderness. Clinicians should maintain a moderate threshold for opening and culturing an incision suspected of being infected and for aspirating any deep fluid collections. Any expressed purulence from within a deep incision consistent with a deep organ space or surgical space infection should be sent for Gram stain and culture. Tissue biopsy and aspiration are recommended methods to obtain infected material.3 In contrast, superficial surgical site infections may be adequately treated with incision, drainage, and local antiseptic care without systemic antibiotic therapy. Superficial incision site swabs are not recommended.3
Suspected Central Nervous System Infection
Assessment
Some of the characteristics that should trigger a clinical assessment for central nervous system (CNS) infection are listed in Table 19.4.
Assessment of a suspected CNS infection must be rapid. The new onset of fever, unexplained altered consciousness, or focal neurologic signs should trigger a diagnostic lumbar puncture unless there are contraindications.7 If the symptoms and signs suggest a focal neurologic process proximal to the spinal cord, imaging studies are indicated before a lumbar puncture is performed to prevent cerebral herniation. If a mass is found, a neurology or neurosurgery consultation should be considered to determine the optimal diagnostic approach. Febrile patients with an intracranial device should have a cerebrospinal fluid (CSF) analysis conducted urgently. Patients with a ventriculostomy who develop the signs or symptoms listed in Table 19.4 and a CSF pleocytosis suggestive of possible microbiological infection should usually have their intracranial catheter removed and the catheter tip cultured.
CSF analysis should involve a cytocentrifuged Gram stain and culture. The intracellular organisms on Gram stain should exclude for the clinician the likelihood of a contaminant. Urgent glucose quantification, protein quantification, and cell count with differential can be used as secondary measures to help determine the presence of infection. However, patients with recent trauma, stroke (hemorrhagic or nonhemorrhagic), neoplasm, or surgery may have derangements of the cell count and protein levels that are not related to infection, and therefore trends or spot checks with these parameters alone may not be helpful. A CSF lactate determination may better assess for infection because an elevated lactate level may provide more specific and sensitive information on infection in the absence of the results of Gram stain and culture.14 Additional tests for infectious pathogens and neoplastic cells should be conducted as indicated by the clinical situation.
Fever and at least one of the following: |
|
|
|
|
|
|
|
|
|
|
Infectious Meningitis
If infectious meningitis is suspected, empiric therapy should be initiated before confirmatory cultures are obtained. Such therapy consists of dexamethasone and empiric antimicrobial pharmacotherapy, such as ceftriaxone and vancomycin. Potential nosocomial bacterial meningitis should be treated with different antibiotic components, usually an agent active against methicillin-resistant Staphylococcus aureus (MRSA) and Pseudomonas aeruginosa. Clinicians should have a low threshold to initiate acyclovir therapy to treat herpes simplex virus infection empirically if the diagnosis is uncertain.
Postoperative CNS infections most commonly include meningitis, subdural empyema, and brain abscess. The incidence of such sequelae after neurologic procedures has been reported in the literature to be anywhere from 0.8 to 7%.15–18 Surgical prophylaxis is most important for S. aureus, Staphylococcus epidermidis, and Propionibacterium acnes, usually with a cephalosporin antibiotic. Patients known to be colonized with MRSA generally should receive prophylaxis with vancomycin. Patients who have been hospitalized and who have received extended β-lactam antibiotics are at risk for infection with Staphylococcus species that are β-lactam–resistant and should be considered candidates for vancomycin prophylaxis. Prophylactic antimicrobial pharmacotherapy before craniotomy has been reported to decrease the incidence of postoperative meningitis by as much as 50%.19