Fever in the ICU

The appearance of a new fever is always a source of concern in a hospitalized patient. This chapter presents the general considerations for a new-onset fever in ICU patients (1), including the potential sources of fever, empiric antibiotic coverage, and the benefits vs. harm of antipyretic therapy.

I. Fever

A. Fever in the ICU

The current guidelines on fever in ICU patients (1) has the following recommendations:

A body temperature ≥38.3°C (101°F) represents a fever, while a lower threshold of 38.0°C (100.4°F) can be used for immunocompromised patients, particularly those with neutropenia.
The most accurate measurements of core body temperature are obtained with thermistor-equipped catheters placed in the pulmonary artery, esophagus, or urinary bladder. Less accurate measurements are obtained with rectal, oral, and tympanic temperature recordings, in that order. The axillary and temporal artery sites are not recommended for temperature measurements.
Comment: Thermistor-equipped urinary bladder cath-eters seem ideal for temperature monitoring in patients who require a bladder catheter (which includes most ICU patients). These devices not only provide reliable
measurements of core body temperature, they also permit continuous temperature monitoring, which has obvious advantages over periodic measurements.

B. Inflammation vs. Infection

Fever is the result of inflammatory cytokines (called endogen-ous pyrogens) that act on the hypothalamus to elevate the body temperature. Any condition that triggers a systemic inflammatory response will produce a fever.

Fever is thus a sign of inflammation, not infection, and about 50% of ICU patients who develop a fever have no apparent infection (2,3).
The severity of a fever does not correlate with the presence or severity of infection. High fevers can be the result of a noninfectious condition such as a drug fever (see later), while fever can be minimal or absent in life-threatening infections (1).
The distinction between inflammation and infection is an important one, not only for the evaluation of fever, but also for curtailing the “knee jerk” response of using antibiotics to treat fever.

II. Noninfectious Sources

Noninfectious sources of fever in the ICU include major surgery, venous thromboembolism, blood transfusions, and drugs.

A. Early Postop Fever

The incidence of fever in the first postoperative day following major surgery is 15–40%, and in most cases, there is no apparent infection (3,4,5). These fevers usually resolve within 24 to 48 hours, and most likely represent an inflammatory response to tissue injury sustained during the surgical procedure.

1. Atelectasis Does Not Cause Fever

There is a longstanding misconception that atelectasis is a common cause of fever in the early postoperative period. One possible source of this misconception is the high incidence of atelectasis in patients who develop a postoperative fever. This is demonstrated in the graph on the left in Figure 35.1 (5), which shows that close to 90% of the patients with fever on the first postoperative day had radiographic evidence of atelectasis. This, however, is not evidence that the atelectasis is the source of fever, as verified by the graph on the right (from the same study), which shows that most (75%) of the patients with atelectasis did not have a fever.

FIGURE 35.1 The relationships between fever and atelectasis in the first postoperative day in 100 consecutive patients who had open heart surgery. The graph on the left shows that most patients with fever had atelectasis, but the graph on the right shows that most patients with atelectasis did not have a fever. Data from Reference 5.

The lack of a causal relationship between atelectasis and fever was demonstrated over 65 years ago in an
animal study that showed the absence of fever after lobar atelectasis was produced by ligation of a mainstem bronchus (6).

2. Malignant Hyperthermia

An uncommon but treatable cause of elevated body temperatures in the immediate postoperative period is malignant hyperthermia, an inherited disorder that produces muscle rigidity, hyperpyrexia (>40°C or >104°F), and rhabdomyolysis in response to halogenated inhalational anesthetics. This disorder is described in Chapter 34, Section II.

B. Venous Thromboembolism

Several groups of patients are at risk for venous thromboembolism, as described in Chapter 4, Section I. Most cases of hospital-acquired deep vein thrombosis are asymptomatic, but acute pulmonary embolism can produce a fever that lasts up to 1 week (7). The diagnostic evaluation for acute pulmonary embolism is described in Chapter 4, Section III.

C. Blood Transfusions

The incidence of febrile, non-hemolytic, transfusion reactions is 1 per 200 transfusions for erythrocytes (see Table 11.3), and 1 per 14 transfusions for platelets (see Table 12.4). These fevers appears during, or up to 6 hours after, the transfusion.

D. Drug Fever

Any drug can trigger a fever (as a hypersensitivity reaction), but ones that are typically implicated in drug fever are listed in Table 35.1.

Drug fever is poorly understood. Over 75% of drug fevers
show no evidence of a hypersensitivity reaction (8).
The onset of the fever varies from a few hours to more than three weeks after the onset of drug therapy (1).
Drug fever can appear as an isolated finding, or it can be accompanied by the other manifestations listed in Table 35.1 (8). These manifestations indicate that drug fever can present as a serious, life-threatening illness.
Suspicion of drug fever usually occurs when there are no other likely sources of fever. When suspected, possible offending drugs should be discontinued. The fever should disappear in 2 to 3 days, but it can persist for up to 7 days (9).

Table 35.1 Drug-Associated Fever in the ICU

Common Offenders	Occasional Offenders	Clinical Findings
Amphotericin	Cimetidine	Rigors (53%)
Cephalosporins	Carbamazepine	Myalgias (25%)
Penicillins	Hydralazine	Leukocytosis (22%)
Phenytoin	Rifampin	Eosinophilia (22%)
Procainamide	Streptokinase	Rash (18%)
Quinidine	Vancomycin	Hypotension (18%)
From Reference 8.

E. Iatrogenic Fever

Faulty thermal regulators in water mattresses and aerosol humidifiers can cause fever by transference (10). It takes only a few minutes to check the temperature settings on heated mattresses and ventilators, but it can take far longer to explain why such a simple cause of fever was overlooked.

III. Nosocomial Infections

The incidence of ICU-acquired infections in medical and surgical ICU patients is shown in Table 35.2 (11). Four infections account for about three-quarters of the infections: pneumonia (mostly ventilator-associated pneumonia), urinary tract infection, bloodstream infection (mostly catheter-related infection), and surgical site infection. Three of these infections are described elsewhere in the book.

Ventilator-associated pneumonias are described in Chapter 16.
Urinary tract infections are described in Chapter 33.
Catheter-related infections are described in Chapter 2, Section III.

The following are the remaining nosocomial infections that deserve mention.

Table 35.2 Nosocomial Infections in Medical and Surgical ICU Patients

Infection	% Total Infections
Infection	Medical ICU	Surgical ICU
Pneumonia	30%	33%
Urinary Tract Infection	30%	18%
Bloodstream Infection	16%	13%
Surgical Site Infection	—	14%
Others	24%	22%
From Reference 11.