Authors: Sarah Dewitt, MD (EM Resident Physician, Virginia Tech-Carilion), Summer Chavez, DO/MPH (EM Resident Physician, Virginia Tech-Carilion), and Jack Perkins, MD (EM Assistant Program Director, Virginia Tech-Carilion) // Edited by: Alex Koyfman, MD (@EMHighAK, EM Attending Physician, UTSW / Parkland Memorial Hospital) and Brit Long, MD (@long_brit)
Case: A 61 y/o male is brought to the Emergency Department (ED) by his family for complaints of dyspnea, subjective fever, and chest pain. He has no past medical history, and his VS include temperature 100.6°F, heart rate of 110, blood pressure 130/80, respiratory rate of 26, and pulse oxygenation 90% on room air. His examination reveals a well-developed male who is in mild distress, but the cardiac and pulmonary examinations are non-contributory. Your initial concerns focus around suspicion of pneumonia, although your resident quickly points out this could also be the presentation of a pulmonary embolus. Does the fever help you differentiate between the two, or would have a lack of fever altered your clinical evaluation? How does the presence of fever shape the differential diagnosis in the ED, and in those patients where an infectious etiology is suspected, should the absence of fever reassure the provider?
Introduction: The pathophysiology of fever
The hypothalamus controls body temperature by balancing inputs from the peripheral nerves that utilize warm/cold receptors in the skin and also analyze the temperature of blood in the surrounding area.1 Fever itself is typically caused by a pyrogen, simply defined as a chemical substance that provokes fever. One such example is exogenous pyrogens, such as those seen in gram-positive bacteria (Staphylococcus aureus enterotoxins) and the superantigens associated with Group A streptococcus and Group B streptococcus microbial infections.1 Many bacteria and fungi can trigger the production and release of cytokines, small proteins that trigger the inflammatory cascade. These cytokines lead to prostaglandin-2 release in peripheral tissues, raising the hypothalamic temperature set point through cAMP release. Central nervous system cytokines are responsible for the hyperpyrexia seen in neurologic trauma and infection.1
Table 1: Differential diagnosis of fever in the ED (note: table not inclusive of all possible causes of fever)
|Infectious Causes of Fever||Non-Infectious Causes of Fever|
|Malignancy (e.g. leukemia, lymphoma, pheochromocytoma)|
• URI, pharyngitis
• Aseptic meningitis
|Autoimmune (e.g. rheumatoid arthritis, systemic lupus erythematosus)|
• Allergic reaction to, or metabolic consequences of drug
• Rocky Mountain Spotted fever
• High external temperatures, or excess exercise
• Subarachnoid hemorrhage
|Embolic vs. Thrombosis vs. Infarction
• Renal infarct
|Blood Transfusion Reaction|
• Munchausen’s vs. Munchausen’s by proxy
How do I proceed to determine if the fever in front of me is from an infectious vs. non-infectious source?
Fever is a common finding in patients presenting to the ED. The differential diagnosis of fever is broad and not limited to infectious etiologies. A key clinical question is deciding whether infection is likely enough to warrant antimicrobial administration. A detailed history and physical exam, the past medical history, current medications (e.g. chemotherapy, glucocorticoids), and recent use of antibiotics may help shape the pre-test probability of an infectious source of fever. However, it is common to need adjunctive laboratory testing or radiographic imaging to further evaluate the source of the fever. Basic testing in the ED often involves a complete blood count (CBC), urinalysis, and a Chest X-ray (CXR). The emergency provider (EP) may deem it appropriate to send a urine culture, blood cultures, and add viral antigen testing in select cases. For example, blood cultures may not be necessary in cases of UTI or pneumonia in patients being discharged; however, they would be useful in cases of severe sepsis or septic shock, as these patients will be admitted and the inpatient team will need this information. It may also be necessary to look further through other serologic testing such as C-reactive protein (CRP), sedimentation rate (ESR), and Procalcitonin.
CRP is an acute phase reactant that becomes elevated in response to inflammatory stimuli. Serum CRP levels surge within 4-6 hours after stimulation, double every 8 hours, and peak after 35-60 hours. It is therefore going to be a more beneficial marker of infection after 12 hours of fever. However, in patients who present with fever at more than 12 hours after onset, it has been shown that serum CRP is elevated significantly in patients with bacterial infections.3 Initially, the measurement of CRP was quantitative and positive in almost all disease states, making it a poor test of choice for diagnosing. Since then, labs have created a specific monoclonal antibody and immunological methods of measurement that have made the CRP test used today very accurate and reproducible. It is quick, and its sensitivity is within 0.04mg/L.4 The value of a single CRP measurement in sepsis diagnosis has been investigated and has been found to be useful in the diagnosis of sepsis.4 In a review published in Intensive Care Med 2002, different studies found that CRP cutoffs between 40-100 mg/L had sensitivity for detecting sepsis between 71-100%, and specificity between 40-85.5%.4 CRP is non-specific, however, and may be elevated in numerous other conditions such as malignancy, rheumatologic disease, and chronic vascular disease among other conditions. It is used most frequently when the provider needs adjunctive information in the search for more unusual sources of infection such as osteomyelitis.
Much like CRP, an ESR is occasionally sent in the ED often as an adjunctive piece of information when searching for more unusual causes of infection such as osteomyelitis or a septic prosthetic joint. It suffers from the same lack of specificity, especially in the older population with numerous comorbid conditions.
Procalcitonin (PCT) is a 116-amino acid peptide that is elevated mainly in response to infectious etiologies. It is much more likely to be elevated in bacterial as opposed to viral infections.5, 6 Serum procalcitonin levels increase significantly in severe systemic infections.5 While PCT has been studied extensively in the inpatient setting to de-escalate antibiotic therapy and to determine utility of empiric antibiotics in COPD exacerbations, its role in the ED has yet to be defined. Currently its use is not widespread, and it has not been well enough studied in the ED to support its use in determining antibiotic deployment in patients with severe sepsis or septic shock.
In sum, it may take a considerable amount of time to make a decision in the ED as to whether the febrile patient truly has an infectious etiology and what antibiotics are most appropriate. In the sickest patient cohort (e.g. shock), empiric antibiotics are clearly indicated when sepsis is high on the differential. However, we are obligated to proceed cautiously in those stable febrile patients in whom infection is not readily apparent. Indiscriminate use of antibiotics is fraught with consequences for the patient and the healthcare system. Keep in mind it may still be unclear after a few hours in the ED as to whether the source of the patient fever is infectious, even after robust use of ancillary testing and a thorough history and physical examination.
How often does sepsis present without a fever?
The SIRS criteria embrace hyperthermia and hypothermia as part of the various parameters to help identify potential sepsis. These criteria have come under scrutiny in recent years, and there is no better example as to the pitfalls of the SIRS criteria than the potential for the septic patient who is afebrile. In the elderly and the immunocompromised patient (e.g. HIV/AIDS, cancer, cirrhosis, diabetes mellitus systemic corticosteroid use, organ transplant, use of immunosuppressant medications), the febrile response to infection might be absent. Some sources report that 20-30% of elderly patients may either remain afebrile or mount a blunted response to infection.8 In fact, some studies postulate that patients who do not mount a febrile response to infection often have more adverse clinical outcomes. A study by Fernandes et al highlighted this risk in elderly patients with bacterial meningitis.9 They found that factors that were independently associated with adverse clinical outcome were older age, absence of fever at ICU admission, and lower GCS score.9
Caterino et al reported that only absence of fever on initial presentation to the ED and initial serum bicarbonate level were independently predictive of patient decompensation after admission to a floor bed as defined by a transfer to the ICU within 48 hours of admission.10 One potential conclusion from this study suggests that the diagnosis of sepsis, subsequent treatment, and assessment of severity of illness are all made more complex in the afebrile patient who truly has sepsis.11 The literature also supports atypical presentations (i.e. afebrile, non-specific complaints such as weakness) of the elderly patients who are bacteremic.12
In sum, do not dismiss the potential for sepsis simply because fever is not present. The older patient and those who are immunocompromised should be expected to present in atypical fashion, and sepsis should be on many differentials regardless of whether fever is documented.
How accurate are oral temperatures compared to other core temperatures?
Oral temperatures are much more convenient that core temperatures in the ED. However, the evidence does not support the accuracy of oral temperatures. A 2015 meta-analysis determined that the accuracy of peripheral temperatures was unacceptable for making clinical decisions.13 Seventy-five studies were included that compared a gold standard core temperature to peripheral temperatures. The authors used ± 0.5 °C as the accepted limit of agreement.13 In patients with hyperthermia, the limits of agreement were -1.44 °C to 1.46 °C, and in hypothermia, -2.07 °C to 1.90 °C.13 The calculated sensitivity was only 64% [95% CI: 55-72%], but specificity performed better at 96% [95% CI: 93-97%].13
In conclusion, if the patient has a fever or is hypothermic by oral temperature, it does not need to be repeated. However, if the patient is normothermic by an oral temperature and the result would change provider management (e.g. sepsis is high on the differential), a core temperature is strongly recommended, either by a rectal thermometer or temperature sensing foley catheter.
Is there any harm in not treating a fever?
While some authors argue that fever places physiologic stress on patients, other data supports that fever enhances the immune system and curbs bacterial growth.14 Additional data from observational studies concluded that a higher early fever was linked to lower mortality rates in ICU patients admitted due to infectious causes.14 The results from a 2015 randomized controlled clinical trial that studied the benefit of treating fever in ICU patients were published in the New England Journal of Medicine.14 Seven hundred ICU patients with fever were randomized to receive either 1000 mg of acetaminophen or placebo.14 There was no statistical significance in number of ICU-free days, 28-day mortality, or 90-day mortality between either group.14 While this one RCT showed no true benefit to treating fever, clinical judgment is important as some patients may be distressed by the fever and appreciate anti-pyretic therapy.
Does the degree of fever help in narrowing the differential in terms of infectious vs. non-infectious etiology?
Fever > 106.7° F is considered to be hyperpyrexia.1 While this can happen in septic patients, this is more common in those with intracranial hemorrhage, neuroleptic malignant syndrome, and heat stroke.1, 15 Additionally, there is a marked difference between hyperthermia and fever. Hyperthermia is treated differently and does not respond to typical anti-pyretics, as there are no pyrogenic molecules.1 Consider what the patient was doing immediately before presentation, such as long exposure in hot temperatures consistent with heat stroke.1 Even rarer are those patients who have pathology of their hypothalamus affecting their set point, such as tumor, trauma or hemorrhage—termed “hypothalamic fever.1”
When you have a patient that presents with undifferentiated hyperpyrexia, think about the following differential to help guide your diagnostic evaluation. A thorough review of the history and current medications is paramount, as serotonin syndrome, heat stroke, sepsis, and thyroid storm have specific treatments and high associated mortality.
|Hyperpyrexia Differential in the ED|
|• Heat Exposure (spectrum of illness)|
|• Neuroleptic Malignant Syndrome|
|• Medication Side Effects|
|• Serotonin Syndrome|
|• Thyroid Storm|
Adapted from: McGugan EA. Hyperpyrexia in the emergency department. Emergency Medicine 2001;13(1):116.
In pediatric patients, it is often assumed that the degree of temperature elevation helps distinguish between a viral and bacterial etiology. However, the literature does not support this assumption.16 In one study, neither maximum temperature, age, or leukocytosis was predictive of an underlying bacterial etiology.16
If I suspect sepsis, should I draw blood cultures in the ED each time the patient spikes a fever > 101?
It is incumbent upon emergency providers to try and obtain blood cultures on patients with severe sepsis or septic shock prior to initiation of antibiotics in a reasonable time frame. However, if the patient ends up in the ED for hours, is there utility in repeating blood cultures every time the patient is febrile? Riedel et al did not find an increased likelihood of documenting bacteremia by employing this technique, and there is always the risk of false positive blood cultures.23 At this time it is recommended to only draw the initial set of blood cultures prior to antibiotic administration. Blood cultures are not recommended in patients who will be discharged, have uncomplicated infectious disease presentations (i.e. cellulitis admitted to a non-ICU setting), or in situations where the results of the cultures will not change management (i.e. community-acquired pneumonia admitted to a non-ICU setting).
How often does fever accompany a PE?
Two historical studies from the 1970’s reported fever in 50% (>37.5°C) and 57.1% (>38°C) of patients with pulmonary embolism respectively.16 The landmark Prospective Investigation of Pulmonary Embolism Diagnosis (PIOPED) study reported a much lower rate of fever in their patients with PE. They found that 14% of patients with fever (≥ 100.0°F) had no identifiable source of that fever other than pulmonary embolus.18 The incidence of pulmonary hemorrhage or infarction was not statistically significant higher in those with fever.18 Calvo-Romero et.al performed a retrospective review of 154 patients with acute pulmonary embolism and found 18.2% had fever (>37°C) without any other known causes.19
Fever is commonly seen in pulmonary embolism, especially low-grade fever, but any clinical significance remains to be determined. A high-grade fever (≥ 101°F) was present in only 6% of the study group in the PIOPED trial.18 In the Calvo-Romero study, 27 of 28 patients had a low-grade fever (temperature between 37°C and 39°C).19 In this study, compared to patients with PE without fever, EKG findings, mortality rates, and chest XR findings were similar.19 In the PIOPED study, 37% of patients who died with a pulmonary embolism had a low-grade fever.18 Higher-grade fevers were more likely to be associated with secondary pneumonitis or widespread pulmonary infarction in the PIOPED group.18 Unfortunately it is not uncommon for the EP to be faced with the dilemma of whether a febrile patient with respiratory complaints has a pulmonary embolus or is septic from pneumonia. The clinician must vigorously examine the history, exam, past medical hisotry, and any ancillary testing (e.g. CXR) to help narrow the differential diagnosis. There are certianly circumstances when advanced imaging such as a CTA may be required to differentiate sepsis from pulmonary embolus, especially if the CXR is non-diagnostic.
- Not all fever is from an infectious source. Keep a broad differential and narrow based on a detailed history, a thorough physical exam, and lab/imaging results.
- Blood cultures are not a routine part of the evaluation of fever and should be deployed in clinical scenarios which are evidence-based (e.g. septic shock) or when the results would affect patient care.
- CRP, PCT, and ESR can be helpful in certain patient care scenarios as adjunctive information when trying to establish the source of a fever as infectious in etiology. Lack of specificity for each of these makes the clinical pre-test probability paramount prior to ordering these tests.
- Not all septic patients have fever! Those without fever have been shown to have worse in-hospital outcomes.
- Oral temperatures can be used for clinical decision making if a fever is documented. However, the poor sensitivity of oral temperatures mandates that a core temperature be obtained if the result would change management. Bottom line: if you suspect sepsis and the patient is afebrile orally, get a core temperature.
- Use your clinical judgment when deciding whether to treat a fever—not all fevers need to be treated with anti-pyretics.
- Do not dismiss the diagnosis of pulmonary embolus because the patient is febrile.
References / Further Reading
- Dinarello CA, Porat R. Fever [Internet]. In: Kasper D, Fauci A, Hauser S, Longo D, Jameson JL, Loscalzo J, editors. Harrison’s Principles of Internal Medicine, 19e. New York, NY: McGraw-Hill Education; 2015 [cited 2016 Oct 29]. Available from: http://mhmedical.com/content.aspx?aid=1120874671
- Morris F, Fletcher A. ABC Of Emergency Differential Diagnosis [Internet]. West Sussex, UK: Blackwell Publishing; [cited 2016 Nov 9]. Available from: https://www.google.com/search?client=safari&rls=en&q=Fletcher_C000.indd+-+ABC+Of+Emergency+Differential+Diagnosis+2009.pdf&ie=UTF-8&oe=UTF-8
- Lee C-C, Hong M-Y, Lee N-Y, Chen P-L, Chang C-M, Ko W-C. Pitfalls in using serum C-reactive protein to predict bacteremia in febrile adults in the ED. Am J Emerg Med 2012;30(4):562–9.
- Povoa P. C-reactive protein: a valuable marker of sepsis. Intensive Care Med 2002;28:235–43.
- Becker KL, Snider R, Nylen ES. Procalcitonin assay in systemic inflammation, infection, and sepsis: clinical utility and limitations. Crit Care Med 2008;36(3):941–52.
- Mat-Nor MB, Ralib A, Abdulah NZ, Pickering JW. The diagnostic ability of procalcitonin and interleukin-6 to differentiate infectious from noninfectious systemic inflammatory response syndrome and to predict mortality. J Crit Care 2016;33:245–51.
- Puskarich MA, Jones AE. Sepsis [Internet]. In: Tintinalli JE, Stapczynski JS, Ma OJ, Yealy DM, Meckler GD, Cline DM, editors. Tintinalli’s Emergency Medicine: A Comprehensive Study Guide, 8e. New York, NY: McGraw-Hill Education; 2016 [cited 2016 Nov 9]. Available from: http://mhmedical.com/content.aspx?aid=1121510693
- Moran D. Infections in the elderly. Top Emerg Med 2003;25(2):174–81.
- Fernandes D, Gonçalves-Pereira J, Janeiro S, Silvestre J, Bento L, Póvoa P. Acute bacterial meningitis in the intensive care unit and risk factors for adverse clinical outcomes: retrospective study. J Crit Care 2014;29(3):347–50.
- Caterino JM, Jalbuena T, Bogucki B. Predictors of acute decompensation after admission in ED patients with sepsis. Am J Emerg Med 2010;28(5):631–6.
- Burlaud A, Mathieu D, Falissard B, Trivalle C. Mortality and bloodstream infections in geriatrics units. Arch Gerontol Geriatr 2010;51(3):e106–9.
- Wester AL, Dunlop O, Melby KK, Dahle UR, Wyller TB. Age-related differences in symptoms, diagnosis and prognosis of bacteremia. BMC Infect Dis 2013;13:346–346.
- Niven DJ, Gaudet JE, Laupland KB, Mrklas KJ, Roberts DJ, Stelfox HT. Accuracy of peripheral thermometers for estimating temperature: a systematic review and meta-analysis. Ann Intern Med 2015;163(10):768–77.
- Young P, Saxena M, Bellomo R, et al. Acetaminophen for Fever in Critically Ill Patients with Suspected Infection. N Engl J Med 2015;373(23):2215–24.
- McGugan EA. Hyperpyrexia in the emergency department. Emerg Med 2001;13(1):116.
- Trautner BW, Caviness AC, Gerlacher GR, Demmler G, Macias CG. Prospective evaluation of the risk of serious bacterial infection in children who present to the emergency department with hyperpyrexia (temperature of 106 degrees F or higher). Pediatrics 2006;118(1):34–40.
- Nucifora G, Badano L, Hysko F, Allocca G, Gianfagna P, Fioretti P. Pulmonary Embolism and Fever. Circulation 2007;115(6):e173–6.
- Stein PD, Afzal A, Henry JW, Villareal CG. Fever in acute pulmonary embolism. Chest 2000;117(1):39–42.
- Calvo-Romero JM, Lima-Rodríguez EM, Pérez-Miranda M, Bureo-Dacal P. Low-grade and high-grade fever at presentation of acute pulmonary embolism. Blood Coagul Fibrinolysis Int J Haemost Thromb 2004;15(4):331–3.