When to suspect nosocomial fever: A practical approach to diagnosis

Adewale Ajumobi, MD, Loma Linda University; David Bland, MD, Loma Linda University Medical Center; and Michael Ing, MD, Loma Linda VA Hospital, Loma Linda, CA

Published Online: April 7, 2008 - 1:48:38 PM (CDT)

Hospital-acquired fever, usually referred to as nosocomial fever, is defined as an oral temperature of 38.1ºC or higher that develops at least 24 hours after admission and is recorded on at least two consecutive occasions in patients with no history of fever in the 7 days preceding its onset. Up to one third of hospitalized patients develop nosocomial fever, and most cases are caused by bacterial infections. Risk factors vary considerably, ranging from ulcers to cerebrovascular disease. Nosocomial fevers can result secondary to infectious, inflammatory, ischemic, or malignant conditions, including urinary tract infections, pneumonia, Clostridium difficile colitis, or bloodstream infections. Nosocomial infection can result in a four-fold increase in fatal outcomes, highlighting the importance of making an immediate and accurate diagnosis to facilitate appropriate treatment.

Fever is a common clinical finding in hospitalized patients of all ages. Fever that develops in hospitalized patients should be evaluated differently from febrile illnesses present on admission. Fever occurring in a hospitalized patient, often referred to as nosocomial fever, usually results from exposure to pyrogens in the hospital environment or from a medical intervention. Nosocomial fever occurs in 2% to 31% of inpatients, and the majority of these fevers are attributable to bacterial infections.^1-3

Nosocomial febrile illnesses are associated with significant expense and an increased risk of mortality. A 1989 study of 123 patients with nosocomial febrile illness found a mortality rate four times greater than for patients in the control group, who had no such fevers.¹ Few studies have evaluated nosocomial fever as a clinical category. This article offers a pragmatic approach to diagnosing hospital-acquired febrile illnesses.

DEFINITIONS

Several definitions of hospital-acquired fever are used throughout the literature. Filice and colleagues define it as an oral temperature of 38.1ºC or higher that occurs in patients at least 24 hours after hospital admission and without any history of fever in the 7 days preceding its onset.¹ Arbo and associates define nosocomial fever as an oral temperature higher than 38.0ºC that occurs at least 48 hours after admission and is recorded on at least two occasions during any 48-hour period.³

For the purpose of our article, nosocomial fever is defined as an oral temperature of 38.1ºC or higher that develops at least 24 hours after admission and is recorded on at least two consecutive occasions in a patient without any history of fever over the 7 days preceding the fever's onset. We selected a 24-hour period because it encompasses patients who are admitted through the short-stay pathway and develop fever as a consequence of a medical intervention or exposure to a hospital pyrogen.

PATHOPHYSIOLOGY

Fever is a manifestation of the release of proinflammatory cytokines-interleukin (IL)-1α, IL-1β IL-4, IL-6, and tumor necrosis factor α-from macrophages, lymphocytes, fibroblasts, epithelial cells, and endothelial cells as a consequence of infection or inflammation.⁴

Infections, particularly bacterial, represent a major cause of nosocomial fever. In one study, nosocomial fevers in patients older than 60 years were generally associated with bacterial infections.³ The following factors were also predictive of bacterial infection as the source of fever:

• Diabetes mellitus;
• Hypertension;
• Hospital stay >10 days before fever onset;
• Maximum temperature >38.7ºC;
• Indwelling bladder catheterization;
• White blood cell count >10 x10⁹/L; and
• More than 75% neutrophils.

RISK FACTORS

Table 1	Risk factors for nosocomial fever13
Alcoholism
Cerebrovascular disease
Congestive heart failure
"Do-not-resuscitate" status
Fecal incontinence
Foot ulcers
Indwelling bladder catheter placement
Indwelling intravenous catheter placement
Malignancy
Number of procedures before febrile illness
Pressure ulcers

Filice and associates studied 123 military veterans (mean age, 66 years) admitted to a medical ward.¹ They found that nosocomial fever occurred most often in patients who had malignancies, particularly non-Hodgkin's lymphoma, chronic lymphocytic leukemia, and acute myelocytic leukemia. In this study, the average length of hospital stay before nosocomial fever onset was 13.6 days. Other risk factors associated with nosocomial fever included foot ulcers, cerebrovascular disease, and placement of indwelling intravenous and urinary catheters.

Arbo and associates studied 100 patients with a mean age of 59.4 years (range, 22-90 years) treated at a university hospital.³ The study showed that a history of alcoholism, lack of a history of angina or myocardial infarction before fever onset, a do-not-resuscitate status, an increased number of procedures performed in the 72 hours preceding the febrile episode, and placement of an indwelling bladder catheter were significantly more likely in patients with nosocomial fever than in their matched controls.³

Another study followed 608 patients who were 65 years or older (mean age, 85.2 years) in an acute care geriatric unit of a university hospital.² Using a multivariate analysis, the study showed that fecal incontinence, congestive heart failure, and pressure ulcers were independent risk factors for nosocomial febrile illnesses. This study also demonstrated that the number of invasive procedures preceding a febrile episode was a significant predictor of infection. The risk factors for nosocomial fever are listed in Table 1.

ETIOLOGY

The causes of nosocomial fever can be grouped into four clinical entities: infections, inflammation, ischemia, and malignancy. Infections constitute the most common cause of nosocomial fever.

Pneumonia and urinary tract infections (UTIs) accounted for 35% of the diagnoses in the study by Filice and associates.¹ In the study by Trivalle and colleagues, 74% of all cases of nosocomial fever were infectious in origin.² Out of 100 patients with nosocomial fever in the study by Arbo and associates, 51 had bacterial infections, 5 had nonbacterial infections, and 25 had fevers with a noninfectious etiology.³ The etiology of the fevers in the remaining 19 patients was not identified.

A list of causes of nosocomial fever is outlined in Table 2. Physicians should be aware of these causes, both common and uncommon, to ensure that patients are properly treated.

VASCULAR CATHETER-RELATED BLOODSTREAM INFECTIONS

A vascular catheter-related bloodstream infection is caused by an organism that has colonized a vascular catheter.⁵ Intravascular catheters can cause fever by different mechanisms, including localized infection or inflammation, systemic infection, and, rarely, through allergic reaction to chemically impregnated catheters.

The risk of catheter-related sepsis varies according to the type or location of the catheter, techniques used for insertion and manipulation, frequency of manipulation, duration the catheter has been in place, and patient population (eg, neonates, burn victims, patients on chronic antibiotic therapy, etc).⁶

In suspected cases of catheter-associated bloodstream infection, the following strategies are recommended:

• Examine the access sites of all intravascular catheters for redness, swelling, excoriation, or exudates;
• Use Gram's stain and culture analysis for all exudates;
• Remember that the absence of local signs of infection does not rule out catheter-related bloodstream infection; and
• Obtain two sets of blood cultures-two percutaneously, or one from the catheter and the other percutaneously.

Based on clinical suspicion and culture results, the intravascular catheter should be removed and the tip cultured. If needed, a new catheter should be inserted at a different site.⁷

Transesophageal echocardiography is recommended in cases of staphylococcal bacteremia to evaluate patients for endocarditis. If evidence of endocarditis is found, other imaging modalities should be used to investigate for septic emboli, such as computed tomography (CT) scanning or magnetic resonance imaging (MRI) of the abdomen, pelvis, and brain.

Obtaining blood cultures appropriately is essential to making a correct diagnosis. Blood cultures should be obtained in accordance with the following guidelines^6,8:

• Site of venipuncture should be cleansed with 2% aqueous chlorhexidine gluconate, 10% povidone-iodine, or with 2% tincture of iodine (recent studies favor 2% chlorhexidine gluconate); and
• The access to an intravascular device and the stopper on the culture bottle should be cleansed with 70% alcohol.

Table 2	Causes of nosocomial fever
Infection Urinary tract infection Pneumonia Bacteremia Skin and soft tissue infection Central nervous system infection Catheter-related gastrointestinal infection Peritonitis Sinusitis Upper respiratory infection Diverticulitis Cholangitis Vascular infection Device-related colitis Tuberculosis Human immunodeficiency virus/Acquired immune deficiency syndrome
Inflammation Aspiration Acute respiratory distress syndrome Arthritis Autoimmune condition Adrenal insufficiency Connective-tissue disorder Drug-induced fever Acalculous cholecystitis Phlebitis Procedure-related pancreatitis Hematoma Thrombosis Sickle-cell crisis Gastrointestinal bleeding Graft-versus-host disease IV contrast reaction
Ischemia Stroke Myocardial infarction Pulmonary embolism Bowel ischemia
Malignancy Leukemia Lymphoma Nonhematologic cancer

The sensitivity of the blood culture hinges on the volume of blood drawn. Whenever possible, at least 10 to 20 mL of blood should be obtained for each culture.⁹ Two blood cultures are sufficient, each drawn by separate venipuncture.¹⁰ If venipuncture is difficult to perform and an intravascular catheter is in place, the second culture can be drawn from the catheter.

Obtain a pair of blood cultures after the initial onset of nosocomial fever. A second pair should be obtained within 24 hours after the first set.¹¹ Thereafter, blood cultures should be obtained based on the pattern of fever and the clinical condition of the patient.

NOSOCOMIAL PNEUMONIA

Nosocomial pneumonia-as evidenced by fever, leukocytosis, purulent sputum production, and the presence of new infiltrate observed on a chest radiograph-represents the second most common nosocomial infection and the most common acquired infection in the intensive care unit (ICU).¹² Risk factors for nosocomial pneumonia include impaired consciousness, intubation, prolonged mechanical ventilation, repeated intubations, supine positioning, chronic antibiotic use, excessive changing of ventilator circuits, neuromuscular disorders, and increased gastric pH.^13-15 The best diagnostic tests for nosocomial pneumonia remain controversial; noninvasive and invasive bronchoscopic tools have comparable diagnostic yields and share similar methodological limitations.¹⁶ If nosocomial pneumonia is suspected, obtaining a chest radiograph is recommended, and, where applicable, sputum should be obtained for Gram's staining and culturing.

NOSOCOMIAL UTIs

UTIs are the most common nosocomial infection outside the ICU.^1-3 The risk factors for this condition include the presence of a urinary catheter, chronic antibiotic use, immobility, stroke, coma, dementia, neuromuscular disorders, and female sex.

Urine samples should be obtained from all patients with a suspected UTI or an unexplained nosocomial febrile illness and should be sent for urinalysis and urine culture. In a patient with an indwelling Foley catheter and a suspected UTI, the catheter should be replaced with a new one, if a catheter is still indicated. The following guidelines are useful in obtaining the correct urine sample for performing urinalysis and urine cultures:

• For patients without a Foley catheter, a midstream clean-catch urine specimen should be collected. In patients with a Foley catheter, the catheter should be clamped and urine aspirated from the port provided for this purpose. Urine should not be collected from the drainage bag.¹¹
• Urine should be transported to the laboratory rapidly to avoid bacterial multiplication. If transport of the urine specimen will be delayed any longer than 1 hour, the specimen should be refrigerated.¹¹
• Routine monitoring of urine or repeated urine studies are unnecessary.¹¹

NOSOCOMIAL DIARRHEA

Nosocomial diarrhea most often develops as a result of Clostridium difficile toxin. Risk factors include the use of antibiotics and immunosuppressants, particularly cytotoxic agents.¹⁷ Almost any antibiotic can trigger C difficile colitis. Common causes include the use of clindamycin, cephalosporins, or penicillins. Although patients with C difficile often have diarrhea, many do not.¹⁷

The following clinical features and laboratory test results have been associated with nosocomial C difficile infection¹⁸:

• Onset of diarrhea 6 or more days after administration of antibiotics;
• Hospital stay longer than 15 days;
• Presence of fecal leukocytes;
• Presence of semi-formed (as opposed to watery) stools; and
• Cephalosporin use.

If C difficile infection is suspected, a complete blood count, including white blood cell differential, and stool tests for fecal leukocytes and C difficile toxin should be undertaken. It is recommended that more than one stool specimen be collected for the C difficile toxin test because the sensitivity of enzyme-immuno-assay toxin for C difficile colitis increases from 72% to 84% in the second specimen.¹⁸ The most cost-effective test for diagnosis is a stool assay for toxin A or B, but tissue culture cytotoxicity assay is the gold standard.¹⁸

DRUG-INDUCED FEVER

Nosocomial fever can result from an adverse reaction to a medication. In 3% to 5% of patients, fever may be the only adverse reaction noted.¹⁹ Drugs can cause fever through several mechanisms, including hypersensitivity, local tissue inflammation, stimulation of heat production, reduction of heat dissipation, alteration of thermoregulation, drug withdrawal, and, rarely, pyrogen contamination of the drug product or delivery device.¹⁹,²⁰ Many drugs can cause fever, and the following should be noted:

• Any drug can cause hypersensitivity.
• Amphotericin B, erythromycin, potassium chloride, and cytotoxics can cause local inflammation.
• Thyroxine may stimulate heat production.
• Anticholinergics and epinephrine can limit heat dissipation.
• Tranquilizers, antihistamines, and phenothiazines can alter thermoregulation.
• Neuroleptics can cause hyperthermia and neuroleptic malignant syndrome.

Drug-induced fever is difficult to identify in hospitalized patients because of confounding comorbid conditions; however, clues to the presence of drug-induced fever include elevation in temperature shortly after a particular drug is administered and cessation of fever once the drug is discontinued.

NOSOCOMIAL FEVER FROM NONOBVIOUS AND UNCOMMON CAUSES

The cause of nosocomial fever is often not apparent, especially when symptoms or obvious physical findings suggestive of a specific illness are absent. In such cases, following the steps described in the diagnostic algorithm can help physicians make the diagnosis (Figure). Physicians should obtain a detailed history and conduct a complete physical examination before performing laboratory or radiologic tests. Special attention should be paid to the patient's surgical history and procedures performed during hospitalization. The patient's medication list also should be reviewed carefully.

Figure—How to evaluate patients with nosocomial fever.

The physical examination should cover all systems and include the skin. The choice of laboratory or imaging studies should be guided by findings from the history and physical examination; for example, the presence of a new heart murmur should raise suspicion for infective endocarditis, and blood cultures and an echocardiogram should be obtained.

Initial laboratory and imaging studies should be noninvasive, inexpensive, and based on the differential diagnosis reached after taking the patient's history and conducting the physical examination. These initial investigations usually include a complete blood count, comprehensive metabolic panel (which includes liver function tests), urinalysis, basic cultures, chest radiographs, and erythrocyte sedimentation rate.

If the initial evaluation does not reveal the cause of nosocomial fever, more expensive and invasive studies can be undertaken. These studies should be based on clinical suspicion and the results of the initial studies and may include ultrasonography, CT scanning, MRI, nuclear medicine scanning, and serology. These modalities, including directed biopsies where indicated, reduce the need for more invasive procedures.²¹ The use of radionuclide scanning, such as technetium Tc 99m, gallium-67, or indium-labeled leukocytes, is warranted for obscure inflammatory or neoplastic conditions that are not diagnosed by conventional imaging studies.²²

CONCLUSION

Nosocomial fever represents a unique clinical entity. Nosocomial UTI is the most common cause of this type of fever in patients hospitalized in the general medicine ward, whereas nosocomial pneumonia is the most common cause of nosocomial fever in ICU patients. Drug-induced fever should always be considered in the differential diagnosis of nosocomial fever.

Taking a detailed history and performing a thorough physical examination are essential and should direct the investigations in pursuit of a diagnosis. Nonobvious and uncommon causes of nosocomial fever may require conducting special laboratory and imaging studies.

PRACTICE POINTS

An oral temperature of 38.1ºC or higher that develops at least 24 hours after admission and is recorded on at least two consecutive occasions constitutes nosocomial fever, if there is no history of fever in the 7 days preceding its onset. With the spread of infections acquired in the hospital and the increased risk of mortality associated with nosocomial fever, early diagnosis is imperative. Fever that develops in the hospital should be evaluated differently from a febrile illness that is present on admission. The most common causes of nosocomial fever are different types of infections; less common causes involve inflammatory, ischemic, and malignant conditions.

SELF-ASSESSMENT TEST

1. Nosocomial fever is defined as:
   
   
   1. Oral temperature of ≥38.1ºC developing at least 24 hours after hospital admission and recorded on two occasions within a 48-hour period.
   2. Oral temperature of ≥38.1ºC developing at least 48 hours after hospital admission and recorded on two occasions within a 24-hour period.
   3. Oral temperature of ≥38.0ºC developing at least 24 hours after hospital admission and recorded on two occasions within a 48-hour period.
   4. Oral temperature of ≥38.0ºC developing at least 48 hours after hospital admission and recorded on two occasions within a 24-hour period.


2. All the following conditions increase the risk for nosocomial fever, except:
   
   
   1. Alcoholism
   2. Do-not-resuscitate status
   3. Depression
   4. Indwelling catheter placement


3. Which of the following is considered the most common cause of nosocomial fever?
   
   
   1. Inflammations
   2. Infections
   3. Malignancies
   4. Ischemic conditions


4. Which of the following statements regarding nosocomial fever is not true?
   
   
   1. An adverse reaction to a medication should always be considered in the differential diagnosis.
   2. The absence of local signs of infection does not rule out catheter-related bloodstream infections.
   3. A chest radiograph is required in all patients with such fever, regardless of the cause.
   4. Taking a detailed history and conducting a thorough physical examination are essential for making the diagnosis.


5. Which of the following is the most common cause of nosocomial diarrhea?
   
   
   1. Helicobacter pylori
   2. Clostridium perfringens
   3. Staphylococcus aureus
   4. Clostridium difficile

(Answers at end of references list)

References

1. Filice GA, Weiler MD, Hughes RA, et al. Nosocomial febrile illnesses in patients on an internal medicine service. Arch Intern Med. 1989;149:319-324.
2. Trivalle C, Chassagne P, Bouaniche M, et al. Nosocomial febrile illness in the elderly: frequency, causes, and risk factors. Arch Intern Med. 1998;158:1560-1565.
3. Arbo MJ, Fine MJ, Hanusa BH, et al. Fever of nosocomial origin: etiology, risk factors, and outcomes. Am J Med. 1993; 95:505-512.
4. Mackowiak PA, Bartlett JG, Borden EC, et al. Concepts of fever: recent advances and lingering dogma. Clin Infect Dis. 1997;25:119-138.
5. Marik PE. Fever in the ICU. Chest. 2000;117:855-869.
6. Centers for Disease Control and Prevention. Guidelines for the Prevention of Intravascular Catheter-Related Infections. Available at: www.cdc.gov/mmwr/preview/mmwrhtml/rr5110a1.htm. Accessed February 6, 2008.
7. Cook D, Randolph A, Kernerman P, et al. Central venous catheter replacement strategies: a systemic review of the literature. Crit Care Med. 1997;25:1417-1424.
8. Strand CL, Wajsbort RR, Sturmann K. Effect of iodophor vs iodine tincture skin preparation on blood culture contamination rate. JAMA. 1993;269:1004-1006.
9. Mermel LA, Maki DG Detection of bacteremia in adults: consequences of culturing an inadequate volume of blood. Ann Intern Med. 1993;119:270-272.
10. Bates DW, Goldman L, Lee TH. Contaminant blood cultures and resource utilization. The true consequences of false-positive results. JAMA. 1991;265:365-369.
11. O'Grady NP, Barie PS, Bartlett J, et al. Practice parameters for evaluating new fever in critically ill adult patients. Crit Care Med. 1998;26:392-408.
12. Lode H, Raffenberg M, Erbes R, et al. Nosocomial pneumonia: epidemiology, pathogenesis, diagnosis, treatment and prevention. Curr Opin Infect Dis. 2000;13:377-384.
13. Torres A, Gatell JM, Aznar E, et al. Re-intubation increases the risk of nosocomial pneumonia in patients needing mechanical ventilation. Am J Resp Crit Care Med. 1995;152:137-141.
14. Holzapfel L, Chevret S, Madinier G, et al. Influence of long-term oro- or nasotracheal intubation on nosocomial maxillary sinusitis and pneumonia: results of a prospective, randomized clinical trial. Crit Care Med. 1993;21:1132-1138.
15. Drakulovic MB, Torres A, Bauer TT, et al. Supine body position as a risk factor for nosocomial pneumonia in mechanically ventilated patients: a randomized trial. Lancet. 1999;354:1851-1858.
16. Ewig S, Bauer T, Torres A. The pulmonary physician in critical care: nosocomial pneumonia. Thorax. 2002;57:366-371.
17. Lai KK, Melvin ZS, Menard MJ, et al. Clostridium difficile-associated diarrhea: epidemiology, risk factors, and infection control. Infect Control Hosp Epidemiol. 1997;18:628-632.
18. Manebe YC, Vinetz JM, Moore RD, et al. Clostridium difficile colitis: an efficient clinical approach to diagnosis. Ann Intern Med. 1995;123:835-840.
19. Roush MK, Nelson KM. Understanding drug-induced febrile reactions. Am Pharm. 1993;NS33:39-42.
20. Hanson MA. Drug fever: remember to consider it in diagnosis. Postgrad Med. 1991;89:167-170, 173.
21. Hirschmann JV. Fever of unknown origin in adults. Clin Infect Dis. 1997;24:291-300.
22. Suga K, Nakagi K, Kuramitsu T, et al. The role of gallium-67 imaging in the detection of foci in recent cases of fever of unknown origin. Ann Nucl Med. 1991;5:35-40.

---

Answers: 1. A; 2. C; 3. B; 4. C; 5. D