Assessment of Fever in the Intensive Care Unit
Is the Answer Just Beyond The Tip of Our Nose?

	ARTICLE

TOP ARTICLE REFERENCES

The emergence of fever during the management of a critically ill patient is an extremely common and challenging problem. In most patients, a broad differential diagnosis exists, including both infectious and noninfectious etiologies. Bedside assessment, including history and physical examination, is hampered by the requirements of intensive care (e.g., sedation and mechanical ventilation) and consequences of the underlying illness (e.g., coma). The inclination to treat aggressively and empirically with broad-spectrum antibiotics is strong, since many clinicians feel these precarious patients cannot withstand progression of a septic process. Yet the adverse effects of empiric therapy in the individual patient and upon the patterns of drug resistance in the unit are real, and thus many advocate a number of specific diagnostic procedures and therapeutic interventions to guide therapy more specifically. Many of these strategies are directed at infectious processes known to arise in the context of critical illness, such as the changing of catheters to reduce the incidence of line-related sepsis or the use of invasive techniques to diagnose ventilator-associated pneumonia (VAP). Some investigators have recommended that evaluation of the febrile intensive care unit (ICU) patient include consideration of nosocomial sinusitis, a complication of mechanical ventilation likely arising from supine positioning and the use of endotracheal, gastric, and feeding tubes (1). Unfortunately, few studies are available that have evaluated such strategies against the endpoints of improved outcome.

In view of this clinical dilemma and the vacuum of information surrounding it, the study by Holzapfel and colleagues in this issue of the American Journal of Respiratory and Critical Care Medicine is of interest to the practicing intensivist (2). These investigators randomized patients in their ICU to receive either standard evaluation of fever occurring during the course of ICU management or a specific diagnostic strategy directed at the possibility of sinusitis. Patients randomized to this specific diagnostic approach underwent sinus computer tomography (CT) scans on Days 4 and 8 after tracheal intubation, and thereafter every seven days if fever ( 38° C) was present. When the CT scan showed an air-fluid level and/or opacification of the maxillary sinus, transnasal puncture was performed for culture, drains were placed, and antibiotics were adjusted in accordance with cultures. The results were striking. Radiographic evidence of sinusitis was observed in 110 of 199 patients in the study group (55%), and 80 patients fulfilled microbiologic criteria for infected maxillary sinusitis, a one-month Kaplan-Meier-estimated occurrence rate of approximately 58%. The percentage of patients receiving antibiotics was increased significantly in the group undergoing imaging of sinuses. Most important, survival at two months was increased in the study group.

This is one of the first investigations suggesting a specific diagnostic pathway to evaluate febrile, critically-ill-patient results in improved outcome. Two important questions arise in considering whether these results are valid and generalizable: (1) is there a plausible mechanism for this substantial clinical benefit? and (2) are there unique features of this patient population that would make the results unlikely to be duplicated in other patient populations?

With regard to mechanistic explanations for the benefit observed, it does not seem likely that sinusitis per se results in progressive and refractory sepsis in all inflicted patients, or that this lethal process would be reversed by the proposed clinical strategy. While sinusitis was a risk factor for bacteremia in this study, the incidence of bacteremia between the study and control groups did not differ. The authors suggest that treatment of sinusitis in the way they have described results in a decreased incidence of ventilator-associated pneumonia (VAP), and hence improved outcome, including improved survival. Several studies have demonstrated an association between nosocomial sinusitis and VAP (3), but the discordance between microbiologic results from sinus and lower airway cultures suggests that in many patients these two processes may simply each arise de novo because of shared risk factors or diminished host defenses for infection in general. In the present study, 37 instances of VAP were noted in the study group (one-month Kaplan-Meier estimate of 34%) versus 51 in the control group (one-month Kaplan-Meier estimate of 47%), a significant difference. However, possible limitations when interpreting this gross observation should be noted. The authors utilized Cox's model to analyze sinusitis as a time-dependent event increasing the risk of VAP and found it did not do so significantly. In addition, no microbiologic studies were done at study initiation to exclude VAP, raising the possibility of group bias. Also, the authors chose protected brush specimens to define VAP microbiologically, a technique that in some studies showed low sensitivity (6, 7). Finally, the authors did not present data that would suggest that the increased incidence of VAP in the control group explained excess mortality.

Thus, from the data available the mechanism of benefit of this strategy is uncertain. This is an important point, since the only clear difference between the study and control group was an increased use of antibiotics in the study population, suggesting that altered protocols for empiric antibiotic use could result in similar outcomes, even without the benefit of sinus imaging and subsequent sinus cultures. Indeed, without knowing in detail the pattern of empiric antiobiotic use in this ICU, it is difficult to determine any benefit of this approach elsewhere.

Were there unique features of this patient population that merit consideration? By the description of the population, there were relatively few trauma patients and very few immunocompromised patients, conditions that are much more prominent in other units and that certainly have the potential to influence the causes of fever and infection. The fraction of patients with chronic obstructive pulmonary disease was high (approximately 25% in both groups), perhaps increasing the predisposition to rhinosinusitis. Most important to the general applicability of these results, however, is that all endotracheal tubes and nasogastric tubes in both study and control patients were placed nasally. Two well conducted prospective clinical trials have suggested that the incidence of nosocomial sinusitis is significantly lower when endotracheal intubation is performed via an orotracheal as opposed to nasotracheal route (3, 5). This point is crucial to interpretation of these results, since a much lower incidence of sinusitis would presumably render a concerted search for its existence largely useless.

Thus, caution must be applied to interpretation of the results of this study. Neither the mechanism for benefit nor its generalizability seem clear. The authors reported that 270 sinus CT scans were conducted in the 199 study patients, with only 53 patients receiving no CT scan. The costs and logistics of this strategy are substantial. While the authors are to be applauded for evaluating a specific diagnostic strategy in this challenging group of patients, further investigations in patients undergoing orotracheal intubation, which include detailed algorithms for empiric antibiotic use, are necessary before their approach can be of general value. For the time being, the proper diagnostic and therapeutic strategies for the febrile ICU patient remain to be defined.

Professor of Medicine, Anesthesia,and Critical CareUniversity of ChicagoChicago, Illinois

	References

TOP ARTICLE REFERENCES

1. Heffner, J. E.. 1994. Nosocomial sinusitis. Am. J. Respir. Crit. Care Med. 150: 608 [Medline].

2. Holzapfel, L., C. Chastang, J. Bohe, G. Demingeon, et al . 1999. A randomized study assessing the systematic search of maxillary sinusitis in mechanically ventilated patients. Amer. J. Respir. Crit. Care Medicine 159: 695-701 .

3. Salord, F., P. Gaussorgues, J. Marti-Flich, M. Sirodot, et al . 1990. Nosocomial maxillary sinusitis during mechanical ventilation: a prospective comparison of orotracheal versus the nasotracheal route for intubation. Int. Care Med. 16: 390-393 [Medline].

4. Holzapfel, L., S. Chevret, G. Madinier, F. Onen, et al. Incidence of longterm oro- or naso-tracheal intubation on nosocomial maxillary sinusitis and pneumonia: results of a randomized clinical trial (300 patients). Crit. Care Med. 21:1132-1138.

5. Rouby, J. J., P. Laurent, M. Gosnach, E. Cambau, et al . 1994. Risk factors and clinical relevance of nosocomial maxillary sinusitis in the critically ill. Am. J. Respir. Crit. Care Med. 150: 776-783 [Abstract].

6. Papazian, L., P. Thomas, L. Garbe, I. Guignon, et al . 1995. Bronchoscopic or blind sampling techniques for the diagnosis of ventilator- associated pneumonia. Am. J. Respir. Crit. Care Med. 152: 1982-1991 [Abstract].

7. Marquette, C. H., M. C. Copin, F. Wallet, R. Neviere, et al . 1995. Diagnostic tests for pneumonia in ventilated patients: prospective evaluation of diagnostic accuracy using histology as a diagnostic gold standard. Am. J. Respir. Crit. Care Med. 151: 1878-1880 [Abstract].