tokines supposedly must cross the blood-brain barrier, which is impermeable to them. At least 2 routes have been proposed: active transport by cytokine-specific receptors and message transfer, where the blood-brain barrier is “leaky,” particularly in the organum vasculsum laminae terminalis in the preoptic area. The other possibilities are that circulating cytokines generate, by the cerebral microvasculature or perivascular microglia and meningeal macrophages, a blood-brain barrier–permeable prostaglandin E2, and a direct transmission of the pyrogenic messages via peripheral vagal afferent nerve to the preoptic areas.9
Beneficial Effects of Fever Because fever is another acute phase response, it is considered beneficial to the host by still poorly understood mechanisms. A recent study compared the activation of TNF in normal and febrile range temperatures, showing that the duration of transcriptional activation was markedly reduced in the 39.5°C cells (30 to 60 minutes) relative to the 37°C cells (2 to 4 hours).10 Another study showed that febrile temperatures (39.2 to 39.7°C) had decreased TNF expression, reduced bacterial load, and improved survival in Klebsiella pneumoniae peritonitis.11 A study in mice showed that at febrile range (39 to 40°C) temperatures, the TNF-␣ levels were increased predominantly in the liver, IL-1␤ levels were higher in the lungs, and IL-6 levels were widely increased in multiple organs.1 This demonstrates that the thermal component of fever may directly contribute to shaping the host response by regulating the timing, magnitude, and tissue distribution of cytokine generation during the acute phase response.
IAN KOMENAKA, MD Department of Surgery New York Methodist Hospital Brooklyn, New York
REFERENCES 1. Jiang Q, Detolla L, Singh IS, et al. Exposure to febrile
temperature upregulates expression of pyrogenic cytokines in endotoxin-challenged mice. Am J Physiol. 1999; 276:R1653-R1660. 2. IUPS Thermal Commision. Glossary of terms for thermal
physiology. Second edition. Pflugers Arch. 1987;410:567587. 3. Dellinger EP. Approach to the patient with postoperative
fever. In: Gorbach SL, Bartlett JG, Blacklow NR, eds. Infectious Diseases. 2nd ed. Philadelphia, Pa: WB Saunders; 1998:903-909. 4. Mackowiak PA, Wasserman SS, Levine MM. A critical
appraisal of 98.6 degrees F, the upper limit of normal body temperature, and other legacies of Carl Reinhold August Wunderlich. JAMA. 1992;268:1578-1580. 5. Hammell HT. Regulation of internal body temperature.
Annu Rev Physiol. 1968;30:641. 6. Fong Y, Moldawer LL, Shires GT, et al. The biologic
characteristics of cytokines and their implication in surgical injury. Surg Gynecol Obstet. 1990;170:363-378. 7. Moldawer LL, Gelin J, Schersten T, et al. Circulating
interleukin 1 and tumor necrosis factor during inflammation. Am J Physiol. 1987;253:R922-R928. 8. Kristiansson M, Soop M, Saraste L, et al. Post-operative
Heat Shock Proteins Heat shock proteins are known to participate in cellular repair and protective mechanisms. The mechanism of induction of the heat shock proteins is not completely understood. The increase of temperature at the site of inflammation triggers the activation of heat shock transcription factor, which causes the transcription of heat shock genes, and the synthesis of heat shock protein. These heat shock proteins protect cells against the toxic effects of mediators of inflammation and infection. One proposed theory involves the attenuation of the cytokine response by heat shock proteins. Expression of heat shock factor 1 (HSF-1) reduced activity of the TNF-␣ promoter. This production of HSF-1 is thought to block the transcription of TNF-␣ by binding to its 85-nucleotide proximal promoter sequence or to a sequence called the 5⬘-untranslated region.10 The production of heat shock proteins is also associated with the induction or increased expression of other stress proteins, including heme oxygenase, superoxide dismutase, metallothionens, and ferritin, which also contribute to the protective effects of the stress response.12 CURRENT SURGERY • Volume 58/Number 2 • March/April 2001
circulating cytokine patterns—the influence of infection. Intensive Care Med. 1993;19:395-400. 9. Blatteis CM. The afferent signalling of fever. J Physiol.
2000;526:470. 10. Singh IS, Viscardi RM, Kalvakolanu I, et al. Inhibition of
tumor necrosis factor-alpha transcription in macrophages exposed to febrile range temperature. A possible role for heat shock factor-1 as a negative transcriptional regulator. J Biol Chem. 2000;275:9841-9848. 11. Jiang Q, Cross AS, Singh IS, et al. Febrile core tempera-
ture is essential for optimal host defense in bacterial peritonitis. Infect Immun. 2000;68:1265-1270. 12. Polla BS, Bachelet M, Elia G, et al. Stress proteins in
inflammation. Ann N Y Acad Sci. 1998;851:75-85.
CLINICAL SIGNIFICANCE AND APPROACH Diffuse fecal peritonitis carries with it the highest morbidity and mortality of all forms of peritonitis. The degree of contam187
ination with particulate fecal matter in conjunction with the high bacterial load in the distal colon places the patient at high risk for both early and late infectious complications. Although the likelihood of postoperative infection in this patient is high given his presenting problem, the true incidence of infections in patients with fever in the postoperative period is remarkably low, with only 14% to 27% of unselected febrile patients having proven infection.1-3 More recently, a greater understanding of the physiologic and immunologic response to tissue injury and invasive infection has led to a re-evaluation of the value of fever as a marker of infection. The response to tissue injury—whether traumatic or surgically induced—manifests as tachycardia, leukocytosis, and fever, findings identical to that of infection. This response has been referred to as the systemic inflammatory response syndrome4 and tends to abate within the first 48 hours. This phenomenon, in part, underlies the rationale for the approach to any patient with postoperative pyrexia. In brief, the clinical context and the time of onset of fever provide the surgeon with crucial information as to the likely etiology and suggest the most appropriate course of investigation.
Fever in the Immediate Postoperative Period (0 to 48 Hours) Fever in the immediate postoperative period is usually of little clinical significance and, depending on the clinical scenario, may or may not have infection as the underlying etiology. If infectious, the clinical scenario usually provides adequate explanation for its development. By contrast, there are several unusual causes of immediate postoperative fever that are life threatening and should be considered in the differential diagnosis of early postoperative fever. The most common infectious causes of immediate postoperative fever often relate directly to the patient’s indication for operative intervention. For example, manipulation of infected tissue in the form of debridement or abscess drainage may release pyrogenic substances into the circulation. This phenomenon is undoubtedly the cause of early postoperative fever in the case described. With a laparotomy and Hartmann procedure, ongoing peritoneal contamination has been arrested and peritoneal toilet achieved. Additionally, this patient is receiving antimicrobials directed against gram-negative enteric and anaerobic organisms, the organisms implicated in the manifestations of diffuse peritonitis. Thus, from both a surgical and antimicrobial standpoint, this patient has been appropriately managed and this fever can readily be attributed to his postoperative septic response without the need for any diagnostic tests except for a good physical examination. Rarely, early postoperative fevers may be due to pneumonia, a finding suggested by the right basal infiltrate demonstrated on the chest radiograph performed on the first postoperative day. These early pneumonias may occur as a result of perioperative 188
aspiration or, less likely, may have preceded the development of the acute process that led to hospital admission. The broadspectrum antimicrobials this patient is receiving are adequate to treat an early-onset pneumonia such that no further diagnostic tests are required. Certain wound infections are classic for their early presentation and warrant specific mention. Both clostridial and ␤-hemolytic group streptococcal infections may present within the first 48 hours and tend to run a fulminant course. Thus, the wound of any patient with severe systemic signs of infection during the first 48 hours after an operation should be thoroughly inspected. These rapidly progressive infections require operative debridement, drainage, and appropriate antimicrobial therapy. The wound described here is at low risk for early wound infection as it was left to close by secondary intention. Nevertheless, a necrotizing soft tissue infection may arise if there is a significant degree of fascial contamination and/or ischemia or, rarely, at the site of the colostomy if there was significant contamination in the process of bringing the stoma through the abdominal wall. There are several noninfectious causes of early postoperative fevers. Most are clinically insignificant and are diagnosed on the basis of exclusion. For example, anaesthetic agents have been reported to have a direct effect on altering the thermoregulatory set point through an effect on the hypothalamus.5,6 Febrile reactions to blood or blood products transfused intraoperatively may also explain some cases of immediate postoperative fever. The diagnosis cannot be established by any means, and the effects of these fevers are clinically insignificant. The Myth of Atelectasis Traditionally, teaching states that the most common cause of fever within the first postoperative 48 hours is atelectasis. This is perhaps one of the greatest myths propagated throughout the years, as there are neither plausible mechanisms nor clinical evidence supporting the relationship between atelectasis and fever.7 Thus, although the right basal infiltrate on chest x-ray in the case described most likely represents atelectasis, its clinical significance and its role in the development of postoperative pyrexia is probably nil. Extremely high fevers (⬎41°C) are of more concern and require consideration of 4 classic causes of the immediate postoperative hypermetabolic response: • • • •
pheochromocytoma thyroid storm malignant hyperthermia adrenal crisis
Patients with any one of these conditions are usually profoundly ill. Immediate diagnosis is difficult, often leading to a “shotgun approach” to therapy. Because administration of dantrolene is the sole definitive therapy for malignant hyperthermia, a clinical trial of this agent is often used in conjunction with anesthesiology consultation. The patient under discussion does not appear to be demonstrating the hemodynamic instability or other manifestations of these processes. CURRENT SURGERY • Volume 58/Number 2 • March/April 2001
Change of Antibiotics? In response to this early postoperative fever, the infectious disease consultant advised a change of antibiotics to imipenem and clindamycin and ordered extensive cultures. Imipenem is an excellent choice for empiric therapy of severe intra-abdominal infections. Its spectrum of activity includes anaerobic organisms, thus making the addition of clindamycin redundant. Further, several well-performed randomized controlled trials have demonstrated that combination (ie, third-generation cephalosporin ⫹ metronidazole) empiric antimicrobial therapy has equivalent efficacy to monotherapy in patients with severe intra-abdominal infection. In summary, there simply is no rationale for the change in antimicrobials suggested by the consultant, nor a reason for the extensive series of cultures suggested at this early stage in the patient’s course. An early postoperative fever in this patient is anticipated and simply relates to the degree of peritoneal contamination and infection and its manipulation at the time of operation. Antipyretic Therapy? The patient’s fever did not abate over the ensuing days, leading to the administration of a nonsteroidal anti-inflammatory agent. There is increasing evidence that pyrexia is a response favorable to the host, whereby it may serve to hasten clearance of infection.8 Antipyretics are only indicated when the hypermetabolic response induced by hyperthermia has adverse effects on the patient (eg, in patients with severe head injury in whom it is preferable to lower cerebral oxygen requirements or those with limited cardiac reserve). Additionally, administration of antipyretics precludes using body temperature as a marker of infection.
Late Postoperative Pyrexia (>48 Hours) Fevers occurring later than 48 hours after operation have a greater likelihood of being infectious in etiology and mandate a thorough history and physical exam and selective diagnostic tests. The most common etiologies are easily remembered by considering the four Ws: wound, wind, water, and walk. The urinary tract (water) is a frequent site of postoperative infection, particularly in those who undergo urinary catheterization. These infections typically manifest at days 3 to 5. Catheters should be removed and urine cultures sent at this time. The diagnosis can be readily established with a urinalysis. Intravenous catheter sites (also water) should be inspected for erythema and removed if there is any evidence of phlebitis or local cellulitis. Central venous catheters may require either removal or exchange over a wire to confirm the presence of a catheter-related infection. With both a urinary catheter and almost certainly a central venous catheter, these possibilities should be considered in the case described. Deep venous thromboses (DVTs) (walk) may also cause fever in the postoperative setting. However, the relationship between fevers and DVT is not clear-cut, with many patients having DVT without fevers and most patients with fevers having no evidence of DVT.9 Further, the relative rarity of DVT as CURRENT SURGERY • Volume 58/Number 2 • March/April 2001
compared with infectious causes of fever mandates a search for a DVT only in the presence of clinical signs of thromboembolic disease or when all other studies are negative. Nosocomial pneumonia (wind) is a frequent cause of postoperative fevers in the later postoperative period, particularly in patients who are ventilator dependent. In these intubated and ventilated patients there is no consensus on the most sensitive and specific tests for the diagnosis of nosocomial pneumonia. Further, if the patient is ready to be weaned off the ventilator, it is unlikely that an untreated ventilator-associated pneumonia is responsible for the pyrexia. Nevertheless, a chest radiograph and, at a minimum, sputum cultures are warranted in the febrile, mechanically ventilated patient. By day 5, the patient was still running a “spiking” pyrexia. Although fever patterns have some diagnostic utility in classic infectious disease, they are nonspecific in the hospital setting.10 Thus this intermittent (septic/hectic) fever suggests infection, but its pattern offers no clue as to the underlying cause. Persistent fevers on antimicrobials suggest only 1 of 3 possibilities: 1. The etiology of the fever is not related to infection (eg, DVT, drug fever). 2. The patient has developed a superinfection with organisms resistant to the current antimicrobials. 3. The patient has an infection not amenable to treatment with antibiotics (ie, requires surgical drainage). In the first 2 scenarios, persisting with the current antimicrobials without any additional data makes no sense— either the patient does not require antibiotics or he is receiving the incorrect antimicrobials. In the absence of any additional cultures demonstrating resistant organisms, altering antimicrobial therapy cannot be performed rationally and therefore should not be done. This leaves the onus on the surgeon to identify the third possibility, undrained infection. In this scenario, the most likely cause is the wound, the fourth W. Wound infections, more formally classified as surgical site infections (SSIs), are categorized as either superficial or deep. Superficial SSIs are typically evident at days 3 to 5 but may occasionally present anywhere from weeks to months (rarely) postoperatively. These infections are easily diagnosed by simply examining the wound. Wounds exhibiting signs of infection may simply be opened to allow drainage. Antibiotics are indicated in the case of significant systemic signs of infection or cellulitis or in the presence of a prosthesis. In this case, the patient already has an opened wound, making a superficial SSI unlikely. Deep SSIs are notoriously difficult to diagnose on clinical examination and require supplemental imaging studies. The classic deep SSI is exemplified by a postoperative intra-abdominal abscess, the most likely diagnosis in the case presented. Patients with intra-abdominal abscesses typically fail to progress in the later postoperative period. Unfortunately, persistent fevers without a clinically evident source typically lead to a prolonged course of antimicrobial therapy. This approach is not justified given the relative ease by which imaging studies (eg, 189
computed tomography [CT]) provide the diagnosis. These patients simply require abscess drainage. Therefore, fevers that persist beyond 5 to 7 days without an alternate source require a definitive imaging study to rule out a deep SSI. A simple CT scan (or even a rectal examination) would provide the necessary data to proceed with percutaneous (or per rectal) drainage. An additional 48 hours of antimicrobials is justified to cover the bacteremia associated with abscess drainage. After this time, no further antibiotics are warranted.
3. Galicier C, Richet H. A prospective study of postoperative
fever in a general surgery department. Infect Control. 1985;6:487-490. 4. American College of Chest Physicians/Society of Critical
Care Medicine Consensus Conference: definitions for sepsis and organ failure and guidelines for the use of innovative therapies in sepsis. Crit Care Med. 1992;20:864-874. 5. Lenhardt R, Negishi C, Sessler DI. Perioperative fever.
Acta Anaesthesiol Scand Suppl. 1997;111:325-328.
Other Potential Sources of Postoperative Infection There are several poorly classified sources of infection that may cause fevers in the postoperative period. In most cases, the site will be identified because of a specific complaint on the part of the patient. However, critically ill patients may not provide any clues, and thus it is the responsibility of the attending physician to consider these alternate causes of postoperative fever. Most of these possibilities will be evident on clinical exam or relatively simple imaging studies. Postoperative parotitis manifests as swelling and erythema over the parotid gland. Purulent discharge may be seen exuding from the parotid duct. Staphylococcus aureus is usually the causative organism. Treatment requires appropriate antimicrobial therapy, salivating agents, and, occasionally, incision and drainage. Sinusitis is not an uncommon source of occult fevers, particularly in patients with longstanding nasogastric or nasotracheal tubes. Diagnosis requires demonstrating fluid in the sinuses (CT or plain films) and culture of nasal discharge or fluid obtained by antral puncture. Rarely, pseudomembranous colitis may manifest as fever, although diarrhea and a leukocytosis are usually the pre-eminent findings. Finally, acalculous cholecystitis may present as occult fevers. Physical examination and/or imaging studies usually direct the clinician in the correct direction. GIUSEPPE PAPIA, MD Department of General Surgery University of Toronto Toronto, Canada AVERY B. NATHENS, MD Department of Surgery Harborview Medical Center University of Washington Seattle, Washington
REFERENCES 1. Mellors JW, Kelly JJ, Gusberg RJ, Horwitz SM, Horwitz
RI. A simple index to estimate the likelihood of bacterial infection in patients developing fever after abdominal surgery. Am Surg. 1988;54:558-564. 2. Freischlag J, Busuttil RW. The value of postoperative fe-
ver evaluation. Surgery. 1983;94:358-363. 190
6. Circiumaru B, Baldock G, Cohen J. A prospective study
of fever in the intensive care unit. Intensive Care Med. 1999;25:668-673. 7. Engoren M. Lack of association between atelectasis and
fever. Chest. 1995;107:81-83. 8. Hanson DF. Fever, temperature, and the immune re-
sponse. Ann N Y Acad Sci. 1997;813:453-464. 9. Kazmers A, Groehn H, Meeker C. Do patients with acute
deep vein thrombosis have fever? Am Surg. 2000;66:598601. 10. Cunha BA. The clinical significance of fever patterns. In-
fect Dis Clin North Am. 1996;10:33-43.
EDITORIAL COMMENT Give me the power to produce fever and I will cure all disease. (Hippocrates, 460-377 BC) That fever may be a sign of disease was recognized in ancient times when warmth meant life and coldness signified death. Sanctorio Sanctorius (1561-1636) was the first one to use a thermometer in humans, but Herman Boerhaave (1668-1738) was the one to use it first in clinical medicine. In 300 BC Erasistratus considered that fever and inflammation were identical, and in 1850 an army surgeon called George Zimmerman restated that inflammation causes fever. With the development of a short-stemmed and convenient thermometer in England (1866) and following the publication of temperature charts for surgical patients by Joseph Bell (1837-1911), the thermometer came into general use. Julius von Wagner-Jauregg won the 1917 Nobel Prize in medicine for his discovery of malarial fever therapy for the late stage of syphilis. So when did the obsession with the “bad” fever start? What is the reason for the currently prevailing dogma confusing fever with infection, considering fever as bad, and treating it with a barrage of antibiotics and antipyretics—though for centuries it has been known that fever is a product of inflammation and inflammation may be “good”? As Drs. Papia and Nathens suggested, it should be quite clear that the commonplace, blind, extended antibiotic administration, for as long as fever or leukocytosis is present, should be CURRENT SURGERY • Volume 58/Number 2 • March/April 2001