Fever and Infections in the Neurological Intensive Care Unit

Fever and infection are major concerns in the daily practice of neurocritical care. Recognition of the importance of fever and its implications in neurocritical care patients in particular has surged in recent years because of increased awareness of the potential detrimental effects of elevated temperature on acute brain injury. In experimental models of cerebral ischemia and traumatic brain injury (TBI), even 1 to 2 degrees of hyperthermia appears to have deleterious effects on outcome (1, 2 and 3). The relevance of these findings in humans is underscored by the association of fever with poor functional outcome after ischemic stroke (4,5), intracerebral hemorrhage (ICH) (6), and aneurysmal subarachnoid hemorrhage (SAH) (7). Brain temperature elevations also have been associated with elevated intracranial pressure after SAH and TBI (8). These recent insights suggest that fever in patients with acute brain injury should be prevented whenever possible, and aggressively treated when it does occur (9).

PHYSIOLOGY OF THERMOREGULATION

Body temperature is normally tightly regulated at approximately 37.0°C (98.0°F). The term fever is generally used in clinical practice (as well as in this chapter) when body temperatures exceed 38.3°C (101.0°F). Strictly speaking, however, the term fever specifically denotes temperature elevation resulting from an increased temperature set point. Set point elevation results from local hypothalamic prostaglandin E synthesis, which is stimulated by circulating systemic inflammatory (pyrogens) such as tumor necrosis factor or interferon 1β (10). The mediators often but not always originate from a tissue infection such as pneumonia, meningitis, or bacteremia; conditions associated with the systemic inflammatory response syndrome (SIRS), drug allergy, deep vein thrombosis, and a variety of other processes also can lead to cytokine elevation and fever (10). Vasoconstriction and shivering are the body’s principal means of generating heat when body temperature is lower than the hypothalamic set point. An example of this process occurs during the “chill phase” of a fever; despite having an elevated and rising body temperature, the patient feels cold and shivers because his temperature is below the set point (11).

By contrast, the term hyperthermia (or hyperpyrexia) refers to clinical conditions in which body temperature exceeds the hypothalamic set point (11). This scenario occurs when either an external or internal source of heat production overwhelms the normal heatdissipating mechanisms of the body (vasodilation and sweating). The best example of hyperthermia from an external source is heat stroke; prolonged exercise and neuroleptic malignant syndrome are examples of hyperthermia resulting from endogenous heat production. Distinguishing between fever and hyperthermia may have important implications for treatment. Antipyretic nonsteroidal antiin-flammatory
agents are effective for fever, but not hyperthermia. External cooling, by contrast, can result in shivering when used to treat fever, but not hyperthermia.

CENTRAL FEVER

Although regularly encountered and routinely diagnosed by exclusion in clinical practice, “central fever” remains a highly controversial entity, which may relate in part to difficulty in establishing the diagnosis with certainty. Many experts advocate complete avoidance of the term central fever (12). Regardless, refractory high fever (greater than 42°C) in the immediate aftermath of massive supratentorial or brainstem hemorrhage is well described, and these unusual but remarkable cases support the notion that acute brain injury can cause fever in the absence of systemic inflammation or infection (13, 14 and 15). The highest such fevers we have observed have occurred in the preterminal phase of massive anterior cerebral artery aneurysm rupture, where core temperatures of 106°F were exceeded.

A more liberal definition of central fever might be any temperature elevation that occurs after brain injury for which no other explanation exists despite an exhaustive diagnostic evaluation. Unexplained fever of this type occurs in approximately 15% of hospitalized stroke patients, and these fevers develop earlier than infection-related fevers (16). In a study of 367 neurological intensive care unit (neuro-ICU) patients, 28% of fevers were unexplained, and external ventricular drainage (EVD) for intraventricular hemorrhage (IVH) was associated with fevers of this type (17). This observation and others (6,7) support the notion that IVH can cause de novo central fever in humans, or at least disrupt normal thermoregulation and exacerbate temperature elevations resulting from conventional pyrogenic stimuli. The mechanism by which IVH may alter hypothalamic function remains speculative. In fact, it is unclear whether central fever primarily reflects fever (set point elevation) or hyperthermia (via sympathetically mediated vasoconstriction and anhidrosis) per se. Direct hemotoxic damage to thermoregulatory centers in the preoptic region, interference with tonic inhibitory inputs from the lower midbrain that ordinarily suppress thermogenesis, and stimulation of prostaglandin production leading to temperature set point elevation all have been invoked (18).

INCIDENCE OF FEVER AND INFECTION IN NEUROLOGICAL INTENSIVE CARE UNIT PATIENTS

Prospective clinical epidemiologic studies indicate that fever eventually develops in 25% to 50% of neuro-ICU patients. Kilpatrick and colleagues (19) identified at least one febrile episode in 47% of 428 consecutively admitted neurosurgical ICU patients. In this study, fever was associated with increased ICU stays and cranial (as opposed to spinal) procedures. Commichau and colleagues (17) identified fever in 23% of 387 neuro-ICU patients. Fifty-two percent of fevers were explained by an infection, which is similar to the frequency of infection among febrile medical ICU patients (20), and the predominant source of infection was the lung (82%), which parallels the experience of stroke patients in general (21,22). By contrast, a more diverse spectrum of nosocomial infections occur in febrile medical ICU patients (20). Coma and mechanical ventilation increased the risk of infectious fever in neuro-ICU patients, which likely reflects the high frequency of ventilator-associated pneumonia associated with these risk factors (17). Among specific diagnoses, SAH was the only condition associated with an increased risk of infectious or unexplained fever after controlling for other predictors, suggesting a generalized disturbance of thermoregulation in these patients (17).

Dettenkofer and coworkers (23) calculated the frequency of nosocomial infections in 545 neurosurgical ICU patients using Centers for Disease Control criteria and found an overall incidence of 20.7 per 100 patients, which is well within the range of published data from
medical ICUs. Pneumonia was the most common site of infection (Table 7.1), and E. coli, Enterococci, and S. aureus were the most common pathogens identified by culture.

TABLE 7.1. Site-specific incidence rates for nosocomial infections in neurosurgical intensive care unit patients

		Incidence per 100 patients	Incidence density per 1,000 days (procedure)
Pneumonia		9.0	15.1 (ventilator days)
Urinary tract infection		7.3	8.5 (urinary catheters)
Bloodstream infection		1.0	0.9 (central line days)
Meningitis		1.1	NC
Brain abscess/ventriculitis		0.7	NC
Other^a		1.7	NC
	TOTAL	20.7	NC
NC, not calculated.
^aMost often wound infection, bronchitis, local intravenous cellulitis, and diarrhea.
Data from Dettenkofer M, Ebner W, Hans F-J, et al. Nosocomial infections in a neurosurgery intensive care unit. Acta Neurochir 1999;141:1303-1308.

EVALUATION OF THE FEBRILE NEUROLOGICAL INTENSIVE CARE UNIT PATIENT

The initial task for the neurointensivist when evaluating fever is to identify its cause. Although infection is always the main concern, a large number of noninfectious causes of fever must also be considered, particularly drug-induced ones (Table 7.2). In addition to performing the traditional initial evaluation for infectious fever in hospitalized patients (chest radiograph, urinalysis, and blood, sputum, and urine cultures), the clinician should obtain a recent history from the nursing staff and patient or family, perform a careful physical examination, and review the patient’s current
medications (24). Most causes of noninfectious fever are identified by taking the extra time for these often-neglected “fundamentals” of hospital practice. In doing so, the astute clinician may preclude a potentially serious medical complication, such as pulmonary embolism, or at the very least protect the patient from unnecessary empiric antibiotic therapy.

TABLE 7.2. Noninfectious causes of fever in neurocritical care patients

Condition		Key to diagnosis
Common
	Blood product reaction	Recent transfusion
	Cocaine intoxication	Toxicology screen
	Central fever	Exclusion
	Deep vein thrombosis	Lower extremity Doppler
	Drug fever	Rash, eosinophilia, transaminitis
	Gout or pseudogout	Joint tenderness and erythema
	Postsurgical local tissue injury	Erythema and tenderness with negative cultures
	Pulmonary embolism with infarction	Hypoxemia on room air, chest computed tomography (CT) angiogram
	Retroperitoneal hemorrhage	Hematocrit, abdominal/pelvic CT scan
	Sterile chemical meningitis	Meningismus
	Systemic inflammatory response syndrome	Leukocytosis, tachycardia, tachypnea
Uncommon
	Adrenal insufficiency	History, serum electrolytes
	Bowel ischemia	Abdominal tenderness and rigidity
	Neuroleptic malignant syndrome	Muscle turgor and rigidity
	Malignancy (lymphoma, leukemia)	Complete blood cell count, chest/abdomen CT scan
	Malignant hyperthermia	Anesthetic exposure
	Myocardial infarction	Electrocardiogram
	Pancreatitis	Serum amylase
	Pericarditis	Pericardial rub, electrocardiogram
	Thyrotoxic crisis	Thyroid function tests

Noninfectious Causes of Fever

Some causes of noninfectious fever in neuro-ICU patients deserve special comment. A truncal maculopapular rash, eosinophilia, mild transaminates, or relative bradycardia may be clues to the presence of a drug fever. Phenytoin and β-lactamase antibiotics (penicillins and cephalosporins) are the most common culprits; the diagnosis is confirmed by improvement after stopping the offending agent (25). The aromatic anticonvulsants (phenytoin, carbamazepine, phenobarbital) are almost as common in neurological practice. Both, of course, may be associated with a rash. Deep vein thrombosis occurs in 9% of neuro-ICU patients, and should be excluded with lower extremity Duplex ultrasonography or plethysmography (26). The systemic inflammatory response syndrome may be invoked when fever is associated with tachypnea, tachycardia, or leukocytosis in the absence of a infection; subtle consumption coagulopathy (d-dimer elevation, hypofibrinogenemia, prothrombin time elevation) corroborates its presence (27).

Another often unsuspected cause of fever is the neuroleptic malignant syndrome, a rare, idiosyncratic reaction to treatment with a dopamine receptor blockers (e.g., haloperidol, Thorazine, and rarely, L-dopa and other drugs) (28). It is characterized by generalized muscle rigidity with rhabdomyolysis, fever, altered mental status, dysautonomia, elevated creatine kinase levels, and leukocytosis. In some cases the rigidity is mild and the fever is not easily recognized as part of the syndrome. Treatment includes discontinuation of the offending agent, surface cooling, dantrolene [1 to 10 mg/kg per day intravenously (i.v.) given every 4 to 6 hours], and/or bromocriptine (2.5 to 5 mg every 8 hours).

Subarachnoid hemorrhage, IVH, and posterior fossa surgery can lead to a sterile inflammatory meningitis that is characterized by a progressive increase in CSF white blood cell counts and hypoglycorrhachia. The inflammation is caused by red blood cell breakdown and reabsorption, and is associated with the intrathecal production of proinflammatory cytokines such as tumor necrosis factor, IL-1, and IL-6 (29). In addition to fever, worsening headaches and meningismus are typical, and in some cases mental status changes may occur; these signs usually respond to treatment with dexamethasone. Finally, brief low-grade temperature elevations are common after surgery of any type, and in most cases are not associated with atelectasis and infection. Postoperative fever of this type results from local tissue inflammation and injury at the site of surgery. Although atelectasis is often invoked, no correlation exists between the presence or severity of atelectasis and postoperative fever (30). The temperature elevation is typically low grade, and resolves spontaneously within 72 hours. Nonetheless, chest physical therapy is a reasonable treatment for all febrile patients, especially in the absence of an obvious cause.

Diagnostic Studies

The incidence of nosocomial infection rises substantially after the third hospital day. Within the hospital, patients in ICUs have the highest risk for nosocomial infection because of the high intensity of invasive drains and monitors as well as their relative immobility. The most common hospital-acquired infections include urinary tract infections (especially in patients with Foley catheters), pulmonary infections (particularly in mechanically ventilated patients), vascular-catheter related bloodstream infections, antibiotic-associated Clostridium difficile colitis, and wound infections. Less common are infected decubitus ulcers, nosocomial sinusitis related to nasotracheal intubation,
and acalculous cholecystitis. Beyond obtaining a chest radiograph and urinalysis and culturing blood, sputum, and urine, additional tests and directed cultures should be prompted by findings on examination. White blood cell count and erythrocyte sedimentation rate elevation reflect the presence of systemic inflammation, which may or may not be caused by infection. The Foley catheter should be changed if infection is suspected, and all indwelling central venous catheters should be removed and cultured. The presence of diarrhea should prompt testing for C. difficile toxin. Lumbar puncture should be performed in patients who have recently undergone craniotomy or placement of a ventricular drain, or in patients with a traumatic CSF fistula to rule out meningitis. After spinal surgery, osteomyelitis may be an occult source of fever and is difficult to detect. The sedimentation rate is elevated and magnetic resonance imaging (MRI) shows signal change in the spinal marrow and often in the adjacent disc space, quite unlike the pattern in neoplastic invasion. In specific circumstances, computed tomographic (CT) scans of the sinuses, chest, abdomen, or pelvis may be helpful to rule out sinusitis, empyema, acalculous cholecystitis, bowel ischemia, or other conditions, but these ancillary tests should be guided by the physical examination.

NOSOCOMIAL INFECTIONS

The epidemiology, pathogenesis, and pathophysiology of nosocomial infections have become increasingly well understood over the past 30 years. Although effective means of decreasing the risk of acquiring many of these infections are now available, nosocomial infections contribute significantly to morbidity and mortality, and still cannot be entirely prevented (31). The principles of prevention can best be understood by considering the pathogenetic, anatomic, and microbiologic features involved in the acquisition of infection. In many cases, nosocomial infection arises as the direct result of violation of a local host defense. The organisms that cause infection in each site are usually found at the mucosal or skin surface at which the local defenses have been breached. In many cases, an indwelling tube or catheter provides intraluminal fluid through which organisms can be infused. Acquired infections can be caused by either endogenous organisms that were part of the patient’s own flora, or exogenous organisms transmitted within the hospital.

The acquisition of nosocomial infection is increasingly recognized as a complex phenomenon with many stages, often including a change in surface flora; change in characteristics of mucosal surfaces or bacterial strains leading to enhanced adherence; interactions between adherence to tissue and foreign bodies; ability to grow (or survive) on foreign bodies or altered sites; modification of local host defenses; and alteration of flora by antibiotics. The aspect of these changes most visible to the clinician, because it is reflected in changing bacteriologic reports from clinical isolates and therefore may affect therapy, is the change in the patient’s flora after admission to the hospital. For example, the pharyngeal flora often begin to include increasing numbers of enteric Gram-negative bacilli, and fecal and skin flora also can become altered to include more resistant hospital-associated strains (31). Many of these changes in colonizing flora result from the direct transfer of organisms through the hands of hospital personnel.

Patients with acute TBI, at least theoretically, have defects in the cellular arm of the immune system that further increase the risk of nosocomial infection. Wolach and associates (32) found significant deficiencies in neutrophil superoxide release, immunoglobulin production, and T-cell function in 14 comatose TBI patients studied within 72 hours of injury. The use of corticosteroids to treat cerebral edema related to ICH, TBI, and cerebral infarction is ineffective, further increases the risk of infection, and should be avoided (33).

Respiratory Infection

Epidemiology

Pneumonia is by far the most common nosocomial infection among critically ill neurological
patients. Approximately half of all nosocomial pneumonias are associated with mechanical ventilation, a condition known as ventilator-associated pneumonia (VAP) (34). The rate of pneumonia increased 5- to 20-fold in intubated patients, and increases with the duration of mechanical ventilation (34). The crude risk of developing VAP is about 1% to 3% per day in intubated patients (35).

Berrouane and coworkers (36) studied the incidence of pneumonia in 569 neurosurgical ICU patients over a period of 1 year. Pneumonia developed in 22%, and the risk was highest in mechanically ventilated and comatose TBI patients. The risk of pneumonia was greatest during the first 3 days in the ICU, with a second smaller peak at days 5 and 6. The incidence rate of nosocomial pneumonia among TBI patients in this study (34.2 per 1,000 ventilation days) is among the highest reported in any ICU population.

Pathogenesis and Prevention

Although nosocomial pneumonia can result from bacteremia or direct inoculation via an endotracheal tube or bronchoscope, aspiration of bacteria from the oropharynx or stomach is by far the most common route of infection. Obviously, neurological patients with an impaired swallowing mechanism and cough reflex are at particularly high risk for aspiration. Most of this aspiration results from subclinical “microaspiration” and is not associated with major aspiration events after vomiting or eating. Aspiration of acidic gastric contents may lead promptly to widespread pulmonary infiltrates, which result in large part from the combined effects of physical obstruction by food particles and the chemical effect of an acute acid burn (37). Although some of these patients subsequently become infected, many do not, and the role of prophylactic antibiotics in preventing subsequent infection is not clear (38). Although histamine blockers and proton pump inhibitors can result in gastric colonization with bacteria and potentially increase aspiration risk, their benefits in patients who are mechanically ventilated, coagulopathic, or have a history of peptic ulcer disease probably outweighs this risk (39).

Only gold members can continue reading. Log In or Register to continue