Noninfectious causes of fever in adults

Abstract

Fever is a common clinical sign encountered in hospitalized patients and often represents the cardinal sign of infectious processes. However, a number of noninfectious etiologies causing fever should be considered prior to initiating broad-spectrum antibiotic therapy. Reducing unnecessary antibiotic use is crucial in an era of increasing resistance.

In Brief

Fever is a common clinical sign encountered in hospitalized patients and often represents the cardinal sign of infectious processes. However, a number of noninfectious etiologies causing fever should be considered prior to initiating broad-spectrum antibiotic therapy. Reducing unnecessary antibiotic use is crucial in an era of increasing resistance.

Figure

Fever is both a common presenting symptom and sign identified during routine patient care in a variety of clinical settings. In the outpatient and hospital settings, fever may be observed in up to 30% to 50% of patients, with the highest incidence among those admitted to critical care, which may be the results of many etiologies.^1,2 Fever is also a common complaint in the primary care setting, with approximately one-third of ambulatory care patients reporting fever to their primary care provider (PCP).^1,3

Among hospitalized patients, documentation of fever in most cases prompts the collection of at least one culture. However, the presence of fever is not a predictor of positive culture results. The absence of fever is actually not a reliable predictor for the absence of bacteremia.² Despite this, empiric antibiotic therapy is commonly initiated as a result of this nonspecific sign (pending additional clinical information).

Emerging antibiotic resistance, due to overprescribing practices or poor prescribing practices, is also a major concern. In the United States, antibiotic-resistant pathogens were responsible for 2 million illnesses and 23,000 deaths in 2011.^4-6 The majority of antibiotic prescriptions are written in the outpatient setting. Advanced practice providers (APPs), a group that consists of both NPs and physician assistants, are a growing segment of the PCP workforce. When looking at overall prescribing practices by provider group, APPs are responsible both for the highest number of antibiotic prescriptions and greatest use of broad-spectrum antibiotics.⁵ It is important to recognize that fever is often the result of a noninfectious etiology. The purpose of this article is to discuss common noninfectious causes of fever in both ambulatory care and hospital settings.

Pathophysiology of fever and hyperthermia

Normal body temperature in healthy individuals is typically 98.6° F (37° C), but may vary by 0.9° F (0.5° C) to 1.8° F (1° C) throughout the body and at various anatomic sites.⁷ The hypothalamic thermoregulatory center regulates body temperature, balancing metabolic heat production with heat released through respiration and evaporation. Disruption of thermoregulatory mechanisms impairs the body's ability to maintain a state of normothermia.

Increased body temperature is commonly referred to as pyrexia (also known as fever) or hyperthermia.⁸ Although often used interchangeably, these terms represent two distinct entities. Fever is the result of an increase in the hypothalamic set point, maintained by heat-generating and heat-preserving thermoeffectors through endogenous pyrogenesis and antipyretic pathways.⁸ Neuronal activation leads to vasoconstriction, shunting blood from the periphery to internal organs to conserve heat. Both thermogenesis from muscles and increased hepatic metabolism can contribute to increases in temperature. By comparison, hyperthermia is often a combination of exogenous heat exposure and the loss of the body's capacity to lose heat. In these cases, temperatures will frequently exceed 105.8° F (41° C).²

Fever, most commonly defined as core temperature of 101° F (38.3° C) or higher, is a hallmark clinical sign in many infectious diseases.⁷ Development of a fever is the result of a complex interaction with exogenous pyrogens (typically microbial pathogens) or pyrogenic cytokines, including tumor necrosis factor (TNF)-alpha, interleukin (IL)-1, IL-6, ciliary neurotrophic factor, interferon (IFN)-gamma, and IFN-alpha released by myeloid and endothelial cells, with the fenestrated capillaries comprising the organ vasculosum lamina terminalis in the anterior hypothalamus.^9,10

Table: Frequent noninfectious etiologies of fever^{2,10-12,14,15,22,25-30,32-34,38,39,41,44-48}

Noninfectious etiologies of fever or hyperthermia include autoimmune diseases, trauma, inflammatory conditions, environmental stressors, or drugs and may increase body temperature (see Frequent noninfectious etiologies of fever). Some clinical clues may be identified in patients with fever from a noninfectious origin (see Distinguishing infectious vs. noninfectious causes of fever).

Etiology of noninfectious causes of fever

Rheumatic diseases, including vasculitis, connective tissue diseases, and granulomatous diseases, may be responsible for fever in up to 30% of patients with fever of unknown origin.^11,12 Consequently, pyrogenic cytokines play a direct role in temperature elevation, as described above, while also modifying hepatic protein synthesis, including increasing: C-reactive protein, serum amyloid A, ceruloplasmin, alpha-1-antitrypsin, haptoglobin, fibrinogen, and C1 inhibitors; pyrogenic cytokines also decrease transferrin, albumin, and fibronectin, many of which are tests performed during routine workup.¹²

Initial workup should include a thorough exam of mucous membranes, eyes, lymph nodes, temporal arteries, liver, spleen, and skin in addition to antinuclear antibody and rheumatoid factor testing.¹¹ Although the antinuclear antibody and rheumatoid factor are sensitive, they lack specificity, and are therefore most useful in confirming a clinical diagnosis. Importantly, bacterial, mycobacterial, and fungal cultures must be obtained to rule out infectious etiologies. Diagnostic imaging may be of benefit, particularly in patients with positive rheumatologic testing. Confirming or excluding a diagnosis may require invasive diagnostics with lymph node, tissue, or bone marrow biopsies.^11,13

Intracranial hemorrhage/ischemic stroke. Central fever occurs in up to 70% of patients admitted with neurologic conditions.¹⁴ Alterations in body temperature are due to the loss of hypothalamic control. Central fever is likely in patients with subarachnoid hemorrhages, intraventricular hemorrhages, brain tumors, or traumatic brain injuries but occurs less commonly in ischemic stroke.^2,14 Compared with fevers caused by infections, central fever occurs earlier and lasts longer with higher temperatures.

Table: Distinguishing infectious vs. noninfectious causes of fever^{2,7,9,10,13,15,32-34}

Confirming the diagnosis of central fever remains a challenge.¹⁴ In the setting of negative cultures, chest radiographs, and fever onset within 3 days, the probability of central fever is high. Empirical antimicrobial therapy is often initiated early in these patients but should promptly be discontinued if the cultures and chest radiographs are negative. Treatment should be directed at lowering body temperature through intravascular or surface cooling devices.

Thomboembolism. Venous thromboembolism (VTE) has been identified as a noninfectious cause of fever.¹⁵ Vascular damage initiates the coagulation process.¹⁶ Recruitment of inflammatory cells, growth factors, and cytokines, including ILs, TNF-alpha, and IFN-gamma, soon follows.¹⁷ Although the precise mechanism of fever has not been described, fever associated with a pulmonary embolism (PE) and deep vein thrombosis (DVT) are likely the result of a complex interaction between direct vascular irritation, inflammatory cytokines, tissue necrosis, and hemorrhage.¹⁸

Most often, fever is due to a PE but has also been observed with DVT.¹⁸ Signs and symptoms are commonly observed in these patients.¹⁹ Typically, fever associated with VTE is low grade (less than 100.9° F [38.3° C]) but rarely greater than 102° F (38.9° C) and occurs for a short duration. Rates of fever are similar in patients with or without pulmonary hemorrhage or infarction.²⁰ Initiation of anticoagulation resolves fever, but higher mortalities persist compared with those without fever.

Aspiration pneumonitis. Inhalation of oropharyngeal or gastric contents into the lungs may lead to a variety of syndromes, including chemical pneumonitis, primary bacterial aspiration pneumonia, or secondary bacterial infection following chemical pneumonitis.²¹ Large volume (greater than 0.3 mL/kg of body weight or 20 to 25 mL in adults) aspiration of acidic gastric contents (pH less than 2.5) is required to cause chemical pneumonitis.²²

Subsequently, corrosive damage to the alveolar cells occurs followed by neutrophil invasion with concomitant release of proinflammatory cytokines. Due to gastric acid inhibiting bacterial growth, gastric contents are typically sterile. Secondary bacterial infections occur more frequently in patients receiving gastric acid suppressing medications.

Symptom onset, fever, hypoxia, dyspnea, and cough occur rapidly within hours following aspiration.²² Radiographic imaging reveals pulmonary infiltrates in the dependent portions of the lungs depending on patient positioning. Clinical and radiographic improvement occurs quickly (within days) in the majority of patients, whereas some rapidly develop respiratory failure and acute respiratory distress syndrome (ARDS).

Initial improvement is followed by clinical deterioration in a small subset of patients who develop secondary bacterial infection following chemical pneumonitis. Treatment primarily revolves around supportive care involving patient repositioning and pulmonary care; however, empirical antibiotics are frequently started.²¹ Early discontinuation of antibiotics is indicated in patients experiencing rapid clinical improvement and in those with negative or contaminated microbiologic samples.

ARDS. Refractory hypoxemia is the hallmark of ARDS.²³ Increased alveolar permeability and collapse develop following direct or indirect lung injury, including aspiration of gastric contents; lung contusion or trauma; inhalation of toxic substances; drug overdose; transfusions; or infectious pneumonia. The progression of ARDS is characterized by three distinct phases.²⁴ Alveolar edema, in addition to proinflammatory and pyrogenic cytokines within the interstitial and alveolar spaces, recruits neutrophils, which causes further alveolar damage with subsequent development of hyaline membranes in the exudative phase; approximately 7 days later, the proliferative phase begins with development of pulmonary fibrosis and interstitial inflammation.²⁵ Although most patients recover within 21 days, some advance to the fibrotic stage, which is characterized by worsening fibrosis and bullae formation.

An ARDS diagnosis is based on the presence of hypoxemia in the setting of bilateral or alveolar interstitial infiltrates occurring within 7 days of insult (known collectively as the Berlin definition).^24,26 Clinical features observed in ARDS may mimic a variety of other noninfectious diseases, including interstitial pneumonia, eosinophilic pneumonia, bronchiolitis obliterans organizing pneumonia, diffuse alveolar hemorrhage, hypersensitivity pneumonitis, and cardiogenic pulmonary edema.²⁷

Identification of the underlying condition is more vital than incorrectly identifying an infectious pneumonia as the cause of respiratory failure, as no specific drug therapy exists for ARDS. Treatment is supportive with noninvasive or mechanical ventilation with low tidal volumes, fluid conservation, and managing the underlying condition.

Blood product transfusion. Approximately 90% of critically ill patients in ICUs are anemic, and 40% of these receive a transfusion during their hospitalization (accounting for the 15 million units transfused annually).²⁸ Although benefits are obvious, transfusions and therapeutic plasma exchange may result in acute or delayed reactions, including allergic, acute, or delayed hemolytic transfusion reaction; febrile nonhemolytic transfusion reaction; acute lung injury; circulatory overload; graft versus host disease; iron overload; or infections.^28-30 No specific signs or symptoms are able to differentiate transfusions, and therapeutic plasma exchange related reactions from other medical problems.

Transfusion and therapeutic plasma exchange related reactions typically occur within 24 hours but as soon as 15 minutes following completion (depending on severity).^28-30 Fever, chills, pruritus, hives, hypotension, and respiratory distress are observed most frequently. Clinical assessment and patient history in combination with information obtained from blood bank personnel are vital to distinguishing transfusion and therapeutic plasma exchange related reactions from infections.

Acute pancreatitis. Pancreatic enzyme activation leads to autodigestion of the pancreas with resultant inflammation, which is known as acute pancreatitis.³¹ Etiologies of acute pancreatitis are dominated by gallstones and alcohol abuse followed by hypertriglyceridemia, drugs, pancreatic strictures, and rarely, infections.^32-34 Patients generally present with acute onset, epigastric pain that may radiate to the back with accompanying nausea and vomiting. Physical findings may include fever, tachypnea, hypotension, and hypoxemia depending on severity.

Diagnosis of acute pancreatitis requires two of the following three criteria: characteristic abdominal pain, serum amylase or lipase at least three times the upper limit of normal, and inflammatory changes visualized on computed tomography (CT) scan.^31,32 Historically, uncontrolled studies suggested antimicrobial therapy may decrease infection and mortalities, but randomized, double-blind, placebo-controlled trials refute this recommendation.^35-37 Antibiotics should be withheld regardless of type or severity of pancreatic disease unless a confirmed infection is documented.³⁸

Acalculous cholecystitis. Acalculous cholecystitis results from biliary stasis rather than gallstones or biliary sludge occluding the cystic duct as observed in acute cholecystitis.^39,40 Risk factors for developing acalculous cholecystitis include sepsis, trauma, burns, surgery, total parenteral nutrition, and even chronic diseases such as diabetes mellitus, cardiovascular disease, and end-stage renal disease. Gallbladder stasis and distension lead to endothelial injury, inflammation, and subsequent necrosis, which may result in perforation.

Clinical manifestations, such as fever, right upper quadrant pain, nausea, and vomiting, are nonspecific and often indistinguishable from calculous cholecystitis.^38,39 Abdominal ultrasound or CT scan may reveal gallbladder distension without evidence of gallstones. Urgent surgical management via cholecystectomy or percutaneous cholecystostomy is required.

Thyroid storm. Excessive release of thyroid hormones yields a hypermetabolic state known as thyroid storm.⁴¹ Although rare, thyroid storm usually follows a precipitating event, such as surgery, anesthesia, trauma, high doses or withdrawal of thyroid hormone, iodinated contrast, or amiodarone. Signs and symptoms may mimic those observed in infections, such as tachycardia and fever (including hyperpyrexia) due to adrenergic activation.

No universally validated criteria or tools currently exist for diagnosing thyroid storm except for a scoring system developed by Burch and Wartofsky in 1993.^41,42 As a result, clinical features are the focus of diagnosis rather than lab abnormalities, which are comparable to those found in hyperthyroidism. Prompt multifaceted treatment aimed at decreasing adrenergic activation, as well as blocking thyroid hormone production, secretion, and systemic activation, is required due to high mortalities.^41,43

Crystal-induced arthropathies. Gout, a metabolic disease, is characterized by hyperuricemia resulting from diets high in purines, alcohol consumption, diuretic therapy, and impaired renal elimination.^44,45 Approximately 1% to 2% of adults in the developed world and 4% in the United States are affected by gout, with higher rates in males and increasing incidence until 70 years of age.⁴⁵ Acute gouty arthritis may mimic infectious conditions, and therefore, should be considered in the differential diagnosis of individuals presenting with fever, pain, erythema, and swelling—particularly around single or multiple joints.^44,45 Inflammation typically accompanies cellulitis and acute gouty arthritis, which focuses around a joint in the latter condition. Although monoarticular involvement may cause difficulty in distinguishing it from septic arthritis, joint aspiration with subsequent identification of urate crystals by polarized microscopy should confirm gout.⁴⁶

Drug-induced fever. While frequently cited as the cause of fever in clinical practice, drug fever is the true cause of fever in a mere 3% to 5% of cases, with increased risk directly correlated with number of drugs prescribed.⁴⁷ Published cases are extremely heterogeneous and identify numerous drugs as the ultimate cause (but most frequently, antibiotics, antiepileptics, and antiarrhythmics). Although the mechanism is not fully understood, drug fever may be the result of hypersensitivity reactions, including a drug reaction with eosinophilia and systemic symptoms syndrome, thermoregulatory alterations, or idiosyncratic reactions.⁴⁸

Unfortunately, due to a common misconception that all fevers are the result of infectious etiologies, an extensive workup with additional blood cultures, radiographic studies, and a variety of antibiotic agents is frequently pursued, which prolongs hospitalization.⁴⁷ As a result, drug fever is typically a diagnosis of exclusion when fever persists.

Fever commonly occurs 7 to 10 days after initiation of the causative agent but may be drug-specific and occur at any time. No specific fever pattern has been identified, but the degree of pyrexia could vary while presenting as continuous, remittent, intermittent, or hectic.⁴⁸ Nonspecific lab abnormalities, including leukocytosis with eosinophilia, elevated erythrocyte sedimentation rate, or transaminitis, could be present.

If drug fever is suspected, the most likely offending agent should be discontinued first followed by other probable drugs if fevers continue.⁴⁷ Fever cessation should occur within 72 hours but may take weeks if accompanied by cutaneous manifestations.

Treatment of fever and hyperthermia

Increased body temperature may result in physiologic stress.^2,8,9 In patients with hyperthermia, body temperature must be rapidly decreased, whereas the beneficial effects of antipyretics to lower temperature in febrile patients are of considerable debate. Acetaminophen and nonsteroidal anti-inflammatory drugs are efficacious in reducing fever while also limiting the associated physiologic effects but may lead to unwanted adverse reactions.² Although treating fever may improve outcomes, attempts to identify the etiology of the increased body temperature are of most importance.

Conclusions

APPs are a steadily growing segment of the PCP population, and therefore, frequently encounter patients presenting with fever as a cardinal sign in their practice. While fever is a sign that can certainly be indicative of an infectious process, it is important for the APP to consider fever of noninfectious origin. A detailed history and physical exam often provide sufficient clinical clues of noninfectious causes of fever. Suspicion for an alternate diagnosis without clinical signs of sepsis should cause the APP to not initiate antibiotic therapy until a diagnosis is firmly established.