SCIENTIFIC PAPERS
Occult Fever in Surgical Intensive Care Unit Patients Is Seldom Caused by Sinusitis* Karen R. Borman, MD, FACS, Phillip M. Brown, MD, Kimberly K. Mezera, MD, Harish Jhaveri, MD, Dallas, Texas
Febrile intensive care unit (ICU) patients were evaluated prospectively for sinusitis. Of 5 9 8 admissions, 2 6 patients with transnasal cannulas, ICU stays over 4 8 hours, and occult fevers were identi, fled. These 2 6 underwent physical examinations and sinus computed tomographic (CT) scans. Maxfllary centeses and cultures were done in patients with CT abnormalities. Patients with positive scans had nasal tubes removed and received decongestants. Scans were abnormal in 19 ( 7 3 % ) . All patients with major CT changes had positive maxillary taps. Most infections were polymicrobial; enteric bacilli were common. Fever resolved with nonoperative care in 18 ( 9 5 % ) patients; in only 1 patient was fever primarily from sinusitis. Sinus CT scans are often abnormal in ICU patients with occult fevers and transnasal cannulas. Pneumatic otoscopy can serve us a screening tool. Most patients respond to nonoperative management. Remote infections are often present. Although radiographic nosocomial ICU sinusitis is common, it is seldom the sole source of fever or the proximate cause of significant morbidity.
From the Departments of Otorhinolaryngology (PMB), Surgery (KRB, KKM), and Radiology(HJ), The Universityof Texas Scuthwestern MedicalSchool,Dallas, Texas. *This is the Jack A. BarneyResidentAward paper. Requests for reprints should be addressed to Karen R. Borman, MD, Department of Surgery,Universityof Texas SouthwesternMedical Center, 5323 Harry Hines Boulevard,Dallas, Texas 75235-9031. Presented at the 44th Annual Meetingof the SouthwesternSurgical Congress,Soottsdale,Arizona, April 26-29, 1992. 412
Iever, hypermetabolic state, sepsis, and even death of the critically ill patient have been attributed to paraF nasal sinusitis resulting from nasal cannulation [I-3]. Reports of this entity have appeared in increasing numbers since the 1974 account of Arens et al [4]. The relationship between nasal cannulas and sinusitis is well established [3-9]. Relevant pathophysiology includes direct mechanical obstruction of sinus ostia, occlusion of sinus ostia by mucosal edema from transnasal tube trauma, disruption of normal nasal cyclic airflow with secondary changes in sinus outflow, and bacteremia associated with nasotracheal intubation. Prior investigations have suggested that sinusitis develops after nasal cannulation in 2% to 17% of intensive care unit (ICU) patients [3,4], with usual disease onset after 7 days of intubation [1]. Drawbacks of previous reports include retrospective design, sharply limited patient population, vague diagnostic criteria, incomplete physical examination, and delayed or suboptimal radiographic evaluation. Few data are provided about alternative sources of febrile morbidity. In the present study, a systematic diagnostic protocol comprising expanded physical examination, directed computerized tomography (CT), maxillary antral centesis with culture, and ongoing observation for other infections was prospectively applied to a mixed surgical ICU population with possible occult fever from paranasal sinusitis. Various bedside and radiologic diagnostic and therapeutic modalities were compared and were correlated with types of nasal cannulas and clinical outcomes. The significance of sinusitis as a source of fever and infectious complications was evaluated. PATIENTS AND M E T H O D S All patients admitted to the multidisciplinary surgical ICU of Parkland Memorial Hospital in Dallas, Texas, from July 1, 1990, to April 1, 1991, were considered for study. Inclusion criteria were: ICU stay exceeding 48 hours, one or more transnasal cannulas, and fever greater than 37.9~ for which panculture and chest roentgenography were performed. Exclusion criteria included: a history of sinus disease, open wound, pulmonary abnormality of recent onset as evidenced by chest roentgenogram, or a condition precluding safe transport to the CT suite. Patients who were accepted for the study were examined at their bedsides by the researchers for symptoms and signs of sinusitis. Nasal abnormalities were sought by anterior rhinoscopy with nasal speculum. Beginning in September 1991, pneumatic otoscopy was added to the examination. Otoscopy was considered positive if an effusion was present, and no attempt was made to further characterize the effusion. Numbers, locations, and types of indwelling nasal tubes were recorded. Paranasal sinus
THE AMERICAN JOURNAL OF SURGERY VOLUME 164 NOVEMBER 1992
INTENSIVE CARE UNIT SINUSITIS
CT scans without contrast were then performed in a uniform manner. All images were interpreted by a single neuroradiologist, and results were grouped into three categories. Major CT findings included either a large-volume sinus air-fluid level, complete sinus opacification, or both. The presence of mucoperiosteal thickening alone was labeled minor, and CT scans without findings of sinusitis were graded normal. Maxillary antral centeses ipsilateral to CT abnormalities were performed by an investigator in all patients with CT changes. After infiltration local anesthesia, the anterior maxillary wall in the area of the canine fossa was prepared with iodine, and a 16-gauge trocar was introdueed into the maxillary sinus. Any fluid present was aspirated and sent for aerobic and anaerobic cultures; the sinus was not lavaged. Patients with normal sinus CT scans did not undergo centeses. When CT was suggestive of sinusitis, nasal cannulas were removed as soon as clinically feasible, and decongestants were administered. Empiric antibiotics were administered at the discretion of primary physicians pending culture results. Further search for other causes of fever was directed by the primary physician. Patients were followed up until hospital discharge, and positive culture results from all sites were recorded. Sinusitis was said to be a "definite" cause of fever in the presence of major radiographic sinusitis (RS) and the absence of any other febrile source. Patients with minor CT abnormalities and no remote infections were classed as "probable." "Possible" was used to describe those patients with minor CT changes and potential noninfectious reasons for fevers, such as pulmonary contusion with associated atelectasis. Occult fever from sinusitis was considered "unlikely" when RS was present but at least one intercurrent infection was also identified. Fevers in patients with normal CT scans but other specific infections were termed "definite non-sinus," and those without both CT changes and febrile sources were called "cryptic." RESULTS Of 598 patients evaluated for inclusion in the study, 26 (4%) satisfied the entry criteria. Purulent rhinorrhea was present on examination in 3 of 26 (12%) patients. Physical signs of sinusitis were absent in all other patients. Nineteen of 26 patients (73%) had RS, 14 with major and 5 with minor CT abnormalities. The sinus images of seven patients were normal. Major findings were seen on the CT scans of all three patients with rhinorrhea. Pneumatic otoscopy was performed in 18 patients. None of five patients without RS had middle ear effusions. Four of five patients with minor RS had normal otoscopies, whereas one had an effusion. Of eight patients with major RS who underwent otoscopy, seven had middle ear effusions, and one did not. Otoscopy was abnormal in both patients with rhinorrhea who were examined. All 19 patients with RS underwent unilateral maxillary antral centeses, and 15 of 19 (79%) aspirations yielded fluid. Antral taps were positive in all 14 patients with major RS and in 1 of 5 patients with minor RS. Aspirate
TABLE !
Relationship of Nasal Cannula Types to Sinus CT Findings No. of Patients Major RS Minor RS No RS
Tube Type Nasotracheal Nasotracheal alone Nasogastric or nasoenteric Nasogastric or nasoenteric alone
14 6 12 8
11 5 3 2
3 1 2 1
0 0 7 5
RS = radiographic sinusitis; CT = computed tomography.
TABLE II
Intercurrent N o n - S i n u s I n f e c t i o n s Infection Site
RS Present (n = 19)
RS Absent (n = 7)
Pulmonary Urinary Vascular catheters Wound Tracheostome Blood Intra-abdominal abscess
16 4 3 2 2 1 0
2 1 0 1 0 0 1
RS = radiographic sinusitis.
volume exceeded 2 mL in all major RS patients but was only 0.1 mL in the sole minor RS patient with a productive centesis. Organisms were recovered from 10 of 15 (67%) maxillary fluid cultures, all 10 from patients with major RS. Five of 10 positive cultures were polymicrobial, and 6 of 10 grew gram-negative enteric bacilli. Identical flora were identified in sinus and sputum cultures in six (60%) patients with any sinus culture growth. Data relating types of nasal cannulations to presence and severity of RS are shown in Table L Both minor RS and major RS were more common with nasotracheal than nasogastric or naseenteric tubes. Twenty-eight intercurrent non-sinus infections were detected in 16 of 19 (84%) RS patients, and 4 of 7 (57%) with normal CT scans (Table H). One patient developed complex sinusitis manifested by subdural empyema; no other complications of sinusitis occurred. Two RS patients had no distant infections, but other potential sources of fever were identified: pulmonary contusion with atelectasis and perihepatic hematoma, respectively. Causes of fever were never determined in three of seven patients without RS. Febrile sinusitis was, therefore, definite in 1 of 19 (5%) RS patients, probable in none, possible in 2 (10%), and unlikely in 16 (84%). Fever was assigned to definite non-sinus causes in four of seven (57%) patients with normal CT scans and remained cryptic in three (43%). Fever resolved in 18 of 19 (95%) RS patients on a treatment regimen of transnasal cannula removal, decongestants, and antibiotics. One patient underwent early
THE AMERICAN JOURNAL OF SURGERY
VOLUME 164
NOVEMBER 1992
413
BORMAN El" AL
operative sinus drainage and subsequent re,operation for evacuation of subdural empyema. She ultimately died of pneumonia after her family decided against further intervention. Fevers resolved with directed antimiorobial therapy in the four patients with normal CT scans and proven non-sinus infections. All three patients with cryptic fevers were afebrile at hospital discharge.
ever, is very common within the general ICU at-risk population when CT is used prospectively for diagnosis. CT scans were abnormal in 73% of our patients and 64% of the group of O'Reilly et al [6]. Hansen et al [7] found RS in 100% of nasotracheally intubated neurosurgical cases, whereas Bowers et al [9] detected RS in 59% of burn victims with nasotracheal tubes. In contrast, prospective plain roentgenograms demonstrated sinusitis in COMMENTS only 34% of the patients in the series by Desmond et al Paranasal sinusitis has been blamed for occult fever [22]. Stauffer et al [23] diagnosed sinusitis in just 8% of and sepsis in the ICU with increasing frequency. Aggres- patients when signs and symptoms of sinusitis or otitis sive search for and treatment of sinusitis in critically ill media were sought by frequent surveillance examinations patients have been advocated [3,6]. Although paranasal by a single investigator. Sinusitis was identified in merely sinusitis is a common disease, it is seldom associated with 2% of cases retrospectively reviewed by Arens et al [4]. major febrile illness unless the sinusitis has extended to Pneumatic otoscopy proved to be a very good, aladjacent tissues. Such complex cases are generally readi- though not perfect, screening tool for the presence of ly identified by accompanying facial edema or eellulitis, major RS, detecting it in 88% of patients. Negative otospurulent rhinorrhea, Pott's puffy tumor, or neurologic copy correctly identified all CT-negative patients and changes, Devastating complications such as subdural em- nearly all minor RS patients. Our data reaffirm the validpyema and cavernous sinus thrombophlebitis are well ity of maxillary centesis as a test of clinical significance, described but occur rarely, usually outside of the setting with productive aspirates in 93% of major RS taps and of disease confined to the sinuses [10-13]. Prior reports only one minor RS centesis. Furthermore, the positive suggest but do not establish if the natural history of minor RS aspiration yielded only a few drops of fluid, simple, uncomplicated ICU sinusitis is different than that compared with several milliliters from each productive of simple community-acquired sinusitis or why it should major RS tap. be different. Previous studies, however, have often been Our data confirm prior reports of nasotracheal inturetrospective [4], limited to specific diagnostic subsets bation as a substantial risk factor for the development of such as neurotrauma [3,14], or used unclear diagnostic ICU sinusitis [3,14]. RS was present in 100% of our criteria and/or suboptimal diagnostic methods [15,16]. patients with nasotracheal cannulas but in only 42% of CT has emerged as a very sensitive and thorough those without them. As suggested by Desmond et al [22], diagnostic modality for use in the paranasal sinuses [17]. nasogastric and/or nasoentcric tubes are less commonly, The improved sensitivity and specificity of CT over stan- although still reasonably often, associated with RS, afdard radiographs in diagnosing sinusitis are well estab- fecting 43% of our patients with nasogastric tubes alone. lished [6], particularly when compared with the limited RS is more frequent when both nares are cannulated, views and poor quality of most ICU bedside sinus plain increasing from 62% with unilateral intubation to 90% for films. Maxillary antral eentesis has previously served as bilateral tubes. the test of clinical significance of sinus findings from plain Community-acquired sinusitis is generally monomiradiographs [18,19], physical examination [19,20], and crobial, predominantly Haemophilus influenzae, pneutransillumination [18,19]. Our study prospectively com- mococcus, Staphylococcus epidermidis, or ~/-hemolytic bines CT as the diagnostic criterion of sinusitis with max- streptococcus [24]. ICU sinusitis, however, is often polyillary centesis after abnormal CT scan as an index of microbial, and gram-negative enteric bacilli are frequentsinusitis severity and as a source of microbiologic materi- ly cultured [3,14-16,25]. Our data support previous studal. We added pneumatic otoscopy to our bedside exami- ies, 50% of our positive cultures being polymicrobial and nation after the prognostic value of otitis media for sinu- 60% including enteric organisms. Antibiotic therapy disitis in adult trauma patients was reported [21]. Our rected toward ICU sinusitis, therefore, requires broader study population was not limited by the primary diagno- coverage than community sinusitis. Of note are six pasis, but eligibility criteria were designed to include only tients with identical organisms identified by sinus and those patients with apparently occult fevers potentially of sputum cultures, a phenomenon noted before by others sinus origin. No attempt was made to define the exact [3,9,14]. The sinuses may provide a bacterial reservoir for mechanism of sinusitis. the lower respiratory tract, or perhaps the same adherent Of 598 admissions to our multidisciplinary surgical bacterial species are incorporated in and shed from a ICU, only 4% were eligible for study entry, suggesting mucus matrix along the entire length of the endotracheal that the potential number of general ICU patients with tubes [26]. occult fevers solely from paranasal sinusitis is small. This Despite the common occurrence of RS in our study agrees with O'Reilly et al [6], who identified 11 of 263 population, occult fever or other infectious complications (4%) mixed ICU admissions with nasotracheai tubes and could be confidently attributed to sinusitis alone in only unexplained sepsis for further evaluation. The proportion one (5%) case. As might be expected based on the natural of patients at risk may be higher in special diagnostic history of non-ICU sinusitis in this ICU patient, disease subsets, such as head injury [3,14] or burn victims [9]. extended beyond the sinuses. Simple sinusitis possibly Radiographic evidence of paranasal sinusitis, how- contributed significantly to fever in slightly over 10% of 414
THE AMERICAN JOURNAL OF SURGERY
VOLUME 164 NOVEMBER 1992
INTENSIVE CARE UNIT SINUSITIS
subjects. At most, then, 16% of RS patients sustained febrile, infectious morbidity from sinusitis. Definite alternative infectious sources were identified in 84% of patients with RS. Notable were 43% of patients without RS in whom fevers remained cryptic after all investigations, suggesting that truly occult fever is well over twice as common as occult febrile sinusitis in multidisciplinary ICU patients with nasal tubes. Consistent with the limited infectious morbidity of simple ICU sinusitis, nonoperative management of this disease is highly successful, as seen in 95% of our patients. Resolution of sinusitis with medical therapy is also the norm in other series [9,15-16]. Treatment comprises removal or relocation of transnasal cannulas, decongestants, and sinus-directed antibiotics. Antimicrobial choices in our series were primarily made on the basis of treating documented intercurrent infections. What is the most efficient and effective scheme of sinus evaluation and treatment in the ICU patient when fever is initially thought to be occult and transnasal tubes are present? A careful bedside examination including anterior rhinoscopy and pneumatic otoscopy should be performed. Sinusitis should be assumed to be present in all patients with any physical signs of sinusitis and those with otitis media. Medical management should be started while the search for other infectious foci proceeds. If fever persists and an alternative febrile cause is not found, sinus CT scanning allows confirmation of RS, assessment of soft tissue extension, and direction of maxillary centesis for culture. Patients without physical signs or otitis media are unlikely to have major RS, and diligent efforts to identify remote infections must continue. Consideration should be given to topical decongestants and the removal of nasotracheal tubes in this group, since some will have minor RS. CT scan and maxillary centesis, however, should precede initiation of any sinus-directed antibiotics for these patients. Although radiologic evidence of paranasal sinusitis is common in ICU patients with transnasal eannulas and occult fevers, sinusitis is seldom the primary infectious culprit. REFERENCES 1. Deutschman CS, Wilton PB, Sinow J, Thienprasit P, Konstantinides FN, Cerra FB. Paranasal sinusitis: a common complication of nasotracheai intubation in neurosurgical patients. Neurosurgery 1985; 17: 296-9. 2. Kronberg FG, Goodwin WJ Jr. Sinusitis in intensive care unit patients. Laryngoscope 1985; 95: 936-8. 3. Grindlinger GA, Niehoff J, Hughes SL, Humphrey MA, Simpson G. Acute paranasal sinusitis related to nasotracheal intubation of head-injured patients. Crit Care Med 1987; 15: 214-7. 4,. Arens JF, LeJeune FE Jr, Webre DR. Maxillary sinusitis, a complication of nasotracheal intubation. Anesthesiology 1974; 40: 415-6. 5. Berry FA Jr, Blankenbaker WL, Ball CG. A comparison of bacteremia occurring with nasotraeheal and orotracheal intubation. Anesth Analg 1973; 52: 873-6. 6, O'Reilly M J, Reddick E J, Black W, et al. Sepsis from sinusitis in nasotracheally intubated patients. Am J Surg 1984; 147: 601-4. 7. Hansen M, Poulsen MR, Bendixen DK, Hartmann-Andersen F. Incidence of sinusitis in patients with nasotracheal intubation. Br J
Anaesth 1988; 61: 231-2. 8. Stillwell M, Caplan ES. The septic multiple-trauma patient. Infect Dis Clin North Am 1989; 3: 144-83. 9. Bowers BL, Purdue GF, Hunt JL. Paranasal sinusitis in burn patients following nasotracheal intubation. Arch Surg 1991; 126: 411-2. 10. Remmler D, Boles R. Intracranial complications of frontal sinusitis. Laryngoscope 1980; 90: 1814-24. 11. Kaufman DM, Litman N, Miller MH. Sinusitis induced subdural empyema. Neurology (NY) 1983; 33: 123-32. 12. Carter BL, Bankoff MS, Fisk JD. Computed tomographic detection of sinusitis responsible for intracranial and extracranial infections. Radiology 1983; 147: 739-42. 13. Parker GS, Tami TA, Wilson JF, Fetter TW. Intracranial complications of sinusitis. South Med J 1989; 82: 563-8. 14,. Humphrey MA, Simpson GT, Grindlinger GA. Clinical characteristies of nosocomial sinusitis. Ann Otol Rhinol Laryngol 1987; 96: 687-90. 15. Caplan ES, Hoyt NJ. Nosocomial sinusitis. JAMA 1982; 247: 639-41. 16. Bell RM, Page GV, Bynoe RP, Dunham ME, Brill AH. Posttraumatic sinusitis. J Trauma 1988; 28: 923-30. 17. Bilaniuk LT, Zimmerman RA. Computed tomography in evaluation of the paranasal sinuses. Radiol Clin North Am 1982; 20: 51-66. 18. Vuorinen P, Kauppila A, Pulkkinen K. Comparison of results of roentgen examination and puncture and irrigation of the maxillary sinuses. J Laryngol Otol 1962; 76: 359-64. 19. McNeill RA. Comparison of the findings on transillumination, x-ray and lavage of the maxillary sinus. J Laryngol Otol 1963; 77: 1009-13. 20. Hamory BH, Sande MA, Sydnor A Jr, Scale DL, Gwaltney JM Jr. Etiology and antimicrobial therapy of acute maxillary sinusitis. J Infect Dis 1979; 139: 197-202. 21. Christensen L, Schaffer S, Ross SE. Otitis media in adult trauma patients: incidence and clinical significance. J Trauma 1991; 31: 1543-5. 22. Desmond P, Raman R, Idikula J. Effect of nasogastric tubes on the nose and maxillary sinus. Crit Care Med 1991; 19: 509-11. 23. Stauffer JL, Olson DE, Petty TL. Complications and consequences of endotracheal intubation and tracheostomy. Am J Med 1981; 70: 65-76. 24. Ylikoski J, Savolainen S, Jousimies-Somer H. The bacteriology of acute maxillary sinusitis. Otol Rhinol Laryngol 1989; 51: 175-81. 25. Linden BE, Aguilar EA, Allen SJ. Sinusitis in the nasotracheally intubated patient. Arch Otolaryngol Head Neck Surg 1988; 114: 860-1. 26. Sottile FD, Marrie T J, Prough DS, et al. Nosocomial pulmonary infection: possible etiologic significance of bacterial adhesion to endotracheal tubes. Crit Care Med 1986; 14: 265-70.
DISCUSSION Robert E. Condon (Milwaukee, WI): These researchers found only 26 of 598 intensive care unit (ICU) patients who were potential candidates for the development of sinusitis. The selected subset of 26 patients all had computed tomographic (CT) examinations of the sinuses, and most had pneumatic otoseopy. Fourteen patients had opacification or an air-fluid level in the sinus. Ten positive cultures were found by fluid aspiration. Five other patients had only thickening of the sinus lining noted on CT. A small amount of fluid was aspirated from
THE AMERICAN JOURNAL OF SURGERY VOLUME 164 NOVEMBER 1992 415
BORMAN ET AL
only one of these patients, and, on culture, it was found to be negative. Among this total group of 598 patients, only 1 proved to have fever due to sinusitis alone, which certainly justifies the conclusion expressed in the title. The authors have effectively put to rest one of the many imaginary trolls that lurk under the ICU bridge. The etiology of sinusitis in ICU patients has never be~n obscure. Nasotracheal, nasogastric, and other tubes through the nose are the culprit. Dr. Brown, the role of the various tubes wasn't clear to me. Of the 14 patients with major radiographic signs of sinusitis, what kinds of tubes were in place when the diagnosis was made? Another question I have concerns the recommendation for the use of CT in the diagnosis of sinusitis. I don't agree with the contention that most portable sinus films done at the bedside are of poor quality. Quality is a reflection of the training and the supervision of the radiology technicians. Good technicians produce good sinus films, and they're less expensive and do not require moving the patient out of the ICU. Finally, I want to explore what is apparently your practice in the ICU of keeping nasotracheal and nasogastric tubes in place for relatively long periods of time. Nasotracheal intubation is usually used to establish an airway in patients in whom potential cervical spine injuries exist or in whom an awake induction of anesthesia is needed. These nasotracheal tubes are removed or
416
THE AMERICAN JOURNAL OF SURGERY
exchanged for an oral tube shortly after surgery. Nasotracheal tubes are also needed for patients with jaw and facial injuries and are maintained beyond the initial day in those who have had their mouth wired shut. Most such patients are candidates for early elective tracheostomy. In regard to nasogastric tubes, most ICU patients don't need a nasogastric tube after the second or third day, and many don't need them at all. Nasogastric tubes are often used because we don't consider the alternatives. Phillip M. Brown (closing): Dr. Condon, we agree that transnasal cannulas are undoubtedly a major contributing factor in sinusitis. By virtue of the study design, all patients had at least one transnasal cannula. Patients with nasotracheal tubes had the highest frequency of major radiographic sinusitis. We agree that the need for long-term use of nasotracheal and nasogastric tubes in ICU patients is infrequent. Our experience with bedside portable sinus roentgenograms differs from yours. For whatever reasons, the films are often nondiagnostic. In addition, portable studies do not allow evaluation of the sphenoidal and ethmoidal sinuses, which have frequently been involved in our patients. Since many of these surgical ICU patients with occult fevers were also scheduled for abdominal or chest CT scans, sinus CT was Seldom the sole reason for transporting a patient from ICU to radiology.
VOLUME 164
NOVEMBER 1992
Occult Fever in Surgical Intensive Care Unit Patients Is Seldom Caused by Sinusitis* Karen R. Borman, MD, FACS, Phillip M. Brown, MD, Kimberly K. Mezera, MD, Harish Jhaveri, MD, Dallas, Texas
Febrile intensive care unit (ICU) patients were evaluated prospectively for sinusitis. Of 5 9 8 admissions, 2 6 patients with transnasal cannulas, ICU stays over 4 8 hours, and occult fevers were identi, fled. These 2 6 underwent physical examinations and sinus computed tomographic (CT) scans. Maxfllary centeses and cultures were done in patients with CT abnormalities. Patients with positive scans had nasal tubes removed and received decongestants. Scans were abnormal in 19 ( 7 3 % ) . All patients with major CT changes had positive maxillary taps. Most infections were polymicrobial; enteric bacilli were common. Fever resolved with nonoperative care in 18 ( 9 5 % ) patients; in only 1 patient was fever primarily from sinusitis. Sinus CT scans are often abnormal in ICU patients with occult fevers and transnasal cannulas. Pneumatic otoscopy can serve us a screening tool. Most patients respond to nonoperative management. Remote infections are often present. Although radiographic nosocomial ICU sinusitis is common, it is seldom the sole source of fever or the proximate cause of significant morbidity.
From the Departments of Otorhinolaryngology (PMB), Surgery (KRB, KKM), and Radiology(HJ), The Universityof Texas Scuthwestern MedicalSchool,Dallas, Texas. *This is the Jack A. BarneyResidentAward paper. Requests for reprints should be addressed to Karen R. Borman, MD, Department of Surgery,Universityof Texas SouthwesternMedical Center, 5323 Harry Hines Boulevard,Dallas, Texas 75235-9031. Presented at the 44th Annual Meetingof the SouthwesternSurgical Congress,Soottsdale,Arizona, April 26-29, 1992. 412
Iever, hypermetabolic state, sepsis, and even death of the critically ill patient have been attributed to paraF nasal sinusitis resulting from nasal cannulation [I-3]. Reports of this entity have appeared in increasing numbers since the 1974 account of Arens et al [4]. The relationship between nasal cannulas and sinusitis is well established [3-9]. Relevant pathophysiology includes direct mechanical obstruction of sinus ostia, occlusion of sinus ostia by mucosal edema from transnasal tube trauma, disruption of normal nasal cyclic airflow with secondary changes in sinus outflow, and bacteremia associated with nasotracheal intubation. Prior investigations have suggested that sinusitis develops after nasal cannulation in 2% to 17% of intensive care unit (ICU) patients [3,4], with usual disease onset after 7 days of intubation [1]. Drawbacks of previous reports include retrospective design, sharply limited patient population, vague diagnostic criteria, incomplete physical examination, and delayed or suboptimal radiographic evaluation. Few data are provided about alternative sources of febrile morbidity. In the present study, a systematic diagnostic protocol comprising expanded physical examination, directed computerized tomography (CT), maxillary antral centesis with culture, and ongoing observation for other infections was prospectively applied to a mixed surgical ICU population with possible occult fever from paranasal sinusitis. Various bedside and radiologic diagnostic and therapeutic modalities were compared and were correlated with types of nasal cannulas and clinical outcomes. The significance of sinusitis as a source of fever and infectious complications was evaluated. PATIENTS AND M E T H O D S All patients admitted to the multidisciplinary surgical ICU of Parkland Memorial Hospital in Dallas, Texas, from July 1, 1990, to April 1, 1991, were considered for study. Inclusion criteria were: ICU stay exceeding 48 hours, one or more transnasal cannulas, and fever greater than 37.9~ for which panculture and chest roentgenography were performed. Exclusion criteria included: a history of sinus disease, open wound, pulmonary abnormality of recent onset as evidenced by chest roentgenogram, or a condition precluding safe transport to the CT suite. Patients who were accepted for the study were examined at their bedsides by the researchers for symptoms and signs of sinusitis. Nasal abnormalities were sought by anterior rhinoscopy with nasal speculum. Beginning in September 1991, pneumatic otoscopy was added to the examination. Otoscopy was considered positive if an effusion was present, and no attempt was made to further characterize the effusion. Numbers, locations, and types of indwelling nasal tubes were recorded. Paranasal sinus
THE AMERICAN JOURNAL OF SURGERY VOLUME 164 NOVEMBER 1992
INTENSIVE CARE UNIT SINUSITIS
CT scans without contrast were then performed in a uniform manner. All images were interpreted by a single neuroradiologist, and results were grouped into three categories. Major CT findings included either a large-volume sinus air-fluid level, complete sinus opacification, or both. The presence of mucoperiosteal thickening alone was labeled minor, and CT scans without findings of sinusitis were graded normal. Maxillary antral centeses ipsilateral to CT abnormalities were performed by an investigator in all patients with CT changes. After infiltration local anesthesia, the anterior maxillary wall in the area of the canine fossa was prepared with iodine, and a 16-gauge trocar was introdueed into the maxillary sinus. Any fluid present was aspirated and sent for aerobic and anaerobic cultures; the sinus was not lavaged. Patients with normal sinus CT scans did not undergo centeses. When CT was suggestive of sinusitis, nasal cannulas were removed as soon as clinically feasible, and decongestants were administered. Empiric antibiotics were administered at the discretion of primary physicians pending culture results. Further search for other causes of fever was directed by the primary physician. Patients were followed up until hospital discharge, and positive culture results from all sites were recorded. Sinusitis was said to be a "definite" cause of fever in the presence of major radiographic sinusitis (RS) and the absence of any other febrile source. Patients with minor CT abnormalities and no remote infections were classed as "probable." "Possible" was used to describe those patients with minor CT changes and potential noninfectious reasons for fevers, such as pulmonary contusion with associated atelectasis. Occult fever from sinusitis was considered "unlikely" when RS was present but at least one intercurrent infection was also identified. Fevers in patients with normal CT scans but other specific infections were termed "definite non-sinus," and those without both CT changes and febrile sources were called "cryptic." RESULTS Of 598 patients evaluated for inclusion in the study, 26 (4%) satisfied the entry criteria. Purulent rhinorrhea was present on examination in 3 of 26 (12%) patients. Physical signs of sinusitis were absent in all other patients. Nineteen of 26 patients (73%) had RS, 14 with major and 5 with minor CT abnormalities. The sinus images of seven patients were normal. Major findings were seen on the CT scans of all three patients with rhinorrhea. Pneumatic otoscopy was performed in 18 patients. None of five patients without RS had middle ear effusions. Four of five patients with minor RS had normal otoscopies, whereas one had an effusion. Of eight patients with major RS who underwent otoscopy, seven had middle ear effusions, and one did not. Otoscopy was abnormal in both patients with rhinorrhea who were examined. All 19 patients with RS underwent unilateral maxillary antral centeses, and 15 of 19 (79%) aspirations yielded fluid. Antral taps were positive in all 14 patients with major RS and in 1 of 5 patients with minor RS. Aspirate
TABLE !
Relationship of Nasal Cannula Types to Sinus CT Findings No. of Patients Major RS Minor RS No RS
Tube Type Nasotracheal Nasotracheal alone Nasogastric or nasoenteric Nasogastric or nasoenteric alone
14 6 12 8
11 5 3 2
3 1 2 1
0 0 7 5
RS = radiographic sinusitis; CT = computed tomography.
TABLE II
Intercurrent N o n - S i n u s I n f e c t i o n s Infection Site
RS Present (n = 19)
RS Absent (n = 7)
Pulmonary Urinary Vascular catheters Wound Tracheostome Blood Intra-abdominal abscess
16 4 3 2 2 1 0
2 1 0 1 0 0 1
RS = radiographic sinusitis.
volume exceeded 2 mL in all major RS patients but was only 0.1 mL in the sole minor RS patient with a productive centesis. Organisms were recovered from 10 of 15 (67%) maxillary fluid cultures, all 10 from patients with major RS. Five of 10 positive cultures were polymicrobial, and 6 of 10 grew gram-negative enteric bacilli. Identical flora were identified in sinus and sputum cultures in six (60%) patients with any sinus culture growth. Data relating types of nasal cannulations to presence and severity of RS are shown in Table L Both minor RS and major RS were more common with nasotracheal than nasogastric or naseenteric tubes. Twenty-eight intercurrent non-sinus infections were detected in 16 of 19 (84%) RS patients, and 4 of 7 (57%) with normal CT scans (Table H). One patient developed complex sinusitis manifested by subdural empyema; no other complications of sinusitis occurred. Two RS patients had no distant infections, but other potential sources of fever were identified: pulmonary contusion with atelectasis and perihepatic hematoma, respectively. Causes of fever were never determined in three of seven patients without RS. Febrile sinusitis was, therefore, definite in 1 of 19 (5%) RS patients, probable in none, possible in 2 (10%), and unlikely in 16 (84%). Fever was assigned to definite non-sinus causes in four of seven (57%) patients with normal CT scans and remained cryptic in three (43%). Fever resolved in 18 of 19 (95%) RS patients on a treatment regimen of transnasal cannula removal, decongestants, and antibiotics. One patient underwent early
THE AMERICAN JOURNAL OF SURGERY
VOLUME 164
NOVEMBER 1992
413
BORMAN El" AL
operative sinus drainage and subsequent re,operation for evacuation of subdural empyema. She ultimately died of pneumonia after her family decided against further intervention. Fevers resolved with directed antimiorobial therapy in the four patients with normal CT scans and proven non-sinus infections. All three patients with cryptic fevers were afebrile at hospital discharge.
ever, is very common within the general ICU at-risk population when CT is used prospectively for diagnosis. CT scans were abnormal in 73% of our patients and 64% of the group of O'Reilly et al [6]. Hansen et al [7] found RS in 100% of nasotracheally intubated neurosurgical cases, whereas Bowers et al [9] detected RS in 59% of burn victims with nasotracheal tubes. In contrast, prospective plain roentgenograms demonstrated sinusitis in COMMENTS only 34% of the patients in the series by Desmond et al Paranasal sinusitis has been blamed for occult fever [22]. Stauffer et al [23] diagnosed sinusitis in just 8% of and sepsis in the ICU with increasing frequency. Aggres- patients when signs and symptoms of sinusitis or otitis sive search for and treatment of sinusitis in critically ill media were sought by frequent surveillance examinations patients have been advocated [3,6]. Although paranasal by a single investigator. Sinusitis was identified in merely sinusitis is a common disease, it is seldom associated with 2% of cases retrospectively reviewed by Arens et al [4]. major febrile illness unless the sinusitis has extended to Pneumatic otoscopy proved to be a very good, aladjacent tissues. Such complex cases are generally readi- though not perfect, screening tool for the presence of ly identified by accompanying facial edema or eellulitis, major RS, detecting it in 88% of patients. Negative otospurulent rhinorrhea, Pott's puffy tumor, or neurologic copy correctly identified all CT-negative patients and changes, Devastating complications such as subdural em- nearly all minor RS patients. Our data reaffirm the validpyema and cavernous sinus thrombophlebitis are well ity of maxillary centesis as a test of clinical significance, described but occur rarely, usually outside of the setting with productive aspirates in 93% of major RS taps and of disease confined to the sinuses [10-13]. Prior reports only one minor RS centesis. Furthermore, the positive suggest but do not establish if the natural history of minor RS aspiration yielded only a few drops of fluid, simple, uncomplicated ICU sinusitis is different than that compared with several milliliters from each productive of simple community-acquired sinusitis or why it should major RS tap. be different. Previous studies, however, have often been Our data confirm prior reports of nasotracheal inturetrospective [4], limited to specific diagnostic subsets bation as a substantial risk factor for the development of such as neurotrauma [3,14], or used unclear diagnostic ICU sinusitis [3,14]. RS was present in 100% of our criteria and/or suboptimal diagnostic methods [15,16]. patients with nasotracheal cannulas but in only 42% of CT has emerged as a very sensitive and thorough those without them. As suggested by Desmond et al [22], diagnostic modality for use in the paranasal sinuses [17]. nasogastric and/or nasoentcric tubes are less commonly, The improved sensitivity and specificity of CT over stan- although still reasonably often, associated with RS, afdard radiographs in diagnosing sinusitis are well estab- fecting 43% of our patients with nasogastric tubes alone. lished [6], particularly when compared with the limited RS is more frequent when both nares are cannulated, views and poor quality of most ICU bedside sinus plain increasing from 62% with unilateral intubation to 90% for films. Maxillary antral eentesis has previously served as bilateral tubes. the test of clinical significance of sinus findings from plain Community-acquired sinusitis is generally monomiradiographs [18,19], physical examination [19,20], and crobial, predominantly Haemophilus influenzae, pneutransillumination [18,19]. Our study prospectively com- mococcus, Staphylococcus epidermidis, or ~/-hemolytic bines CT as the diagnostic criterion of sinusitis with max- streptococcus [24]. ICU sinusitis, however, is often polyillary centesis after abnormal CT scan as an index of microbial, and gram-negative enteric bacilli are frequentsinusitis severity and as a source of microbiologic materi- ly cultured [3,14-16,25]. Our data support previous studal. We added pneumatic otoscopy to our bedside exami- ies, 50% of our positive cultures being polymicrobial and nation after the prognostic value of otitis media for sinu- 60% including enteric organisms. Antibiotic therapy disitis in adult trauma patients was reported [21]. Our rected toward ICU sinusitis, therefore, requires broader study population was not limited by the primary diagno- coverage than community sinusitis. Of note are six pasis, but eligibility criteria were designed to include only tients with identical organisms identified by sinus and those patients with apparently occult fevers potentially of sputum cultures, a phenomenon noted before by others sinus origin. No attempt was made to define the exact [3,9,14]. The sinuses may provide a bacterial reservoir for mechanism of sinusitis. the lower respiratory tract, or perhaps the same adherent Of 598 admissions to our multidisciplinary surgical bacterial species are incorporated in and shed from a ICU, only 4% were eligible for study entry, suggesting mucus matrix along the entire length of the endotracheal that the potential number of general ICU patients with tubes [26]. occult fevers solely from paranasal sinusitis is small. This Despite the common occurrence of RS in our study agrees with O'Reilly et al [6], who identified 11 of 263 population, occult fever or other infectious complications (4%) mixed ICU admissions with nasotracheai tubes and could be confidently attributed to sinusitis alone in only unexplained sepsis for further evaluation. The proportion one (5%) case. As might be expected based on the natural of patients at risk may be higher in special diagnostic history of non-ICU sinusitis in this ICU patient, disease subsets, such as head injury [3,14] or burn victims [9]. extended beyond the sinuses. Simple sinusitis possibly Radiographic evidence of paranasal sinusitis, how- contributed significantly to fever in slightly over 10% of 414
THE AMERICAN JOURNAL OF SURGERY
VOLUME 164 NOVEMBER 1992
INTENSIVE CARE UNIT SINUSITIS
subjects. At most, then, 16% of RS patients sustained febrile, infectious morbidity from sinusitis. Definite alternative infectious sources were identified in 84% of patients with RS. Notable were 43% of patients without RS in whom fevers remained cryptic after all investigations, suggesting that truly occult fever is well over twice as common as occult febrile sinusitis in multidisciplinary ICU patients with nasal tubes. Consistent with the limited infectious morbidity of simple ICU sinusitis, nonoperative management of this disease is highly successful, as seen in 95% of our patients. Resolution of sinusitis with medical therapy is also the norm in other series [9,15-16]. Treatment comprises removal or relocation of transnasal cannulas, decongestants, and sinus-directed antibiotics. Antimicrobial choices in our series were primarily made on the basis of treating documented intercurrent infections. What is the most efficient and effective scheme of sinus evaluation and treatment in the ICU patient when fever is initially thought to be occult and transnasal tubes are present? A careful bedside examination including anterior rhinoscopy and pneumatic otoscopy should be performed. Sinusitis should be assumed to be present in all patients with any physical signs of sinusitis and those with otitis media. Medical management should be started while the search for other infectious foci proceeds. If fever persists and an alternative febrile cause is not found, sinus CT scanning allows confirmation of RS, assessment of soft tissue extension, and direction of maxillary centesis for culture. Patients without physical signs or otitis media are unlikely to have major RS, and diligent efforts to identify remote infections must continue. Consideration should be given to topical decongestants and the removal of nasotracheal tubes in this group, since some will have minor RS. CT scan and maxillary centesis, however, should precede initiation of any sinus-directed antibiotics for these patients. Although radiologic evidence of paranasal sinusitis is common in ICU patients with transnasal eannulas and occult fevers, sinusitis is seldom the primary infectious culprit. REFERENCES 1. Deutschman CS, Wilton PB, Sinow J, Thienprasit P, Konstantinides FN, Cerra FB. Paranasal sinusitis: a common complication of nasotracheai intubation in neurosurgical patients. Neurosurgery 1985; 17: 296-9. 2. Kronberg FG, Goodwin WJ Jr. Sinusitis in intensive care unit patients. Laryngoscope 1985; 95: 936-8. 3. Grindlinger GA, Niehoff J, Hughes SL, Humphrey MA, Simpson G. Acute paranasal sinusitis related to nasotracheal intubation of head-injured patients. Crit Care Med 1987; 15: 214-7. 4,. Arens JF, LeJeune FE Jr, Webre DR. Maxillary sinusitis, a complication of nasotracheal intubation. Anesthesiology 1974; 40: 415-6. 5. Berry FA Jr, Blankenbaker WL, Ball CG. A comparison of bacteremia occurring with nasotraeheal and orotracheal intubation. Anesth Analg 1973; 52: 873-6. 6, O'Reilly M J, Reddick E J, Black W, et al. Sepsis from sinusitis in nasotracheally intubated patients. Am J Surg 1984; 147: 601-4. 7. Hansen M, Poulsen MR, Bendixen DK, Hartmann-Andersen F. Incidence of sinusitis in patients with nasotracheal intubation. Br J
Anaesth 1988; 61: 231-2. 8. Stillwell M, Caplan ES. The septic multiple-trauma patient. Infect Dis Clin North Am 1989; 3: 144-83. 9. Bowers BL, Purdue GF, Hunt JL. Paranasal sinusitis in burn patients following nasotracheal intubation. Arch Surg 1991; 126: 411-2. 10. Remmler D, Boles R. Intracranial complications of frontal sinusitis. Laryngoscope 1980; 90: 1814-24. 11. Kaufman DM, Litman N, Miller MH. Sinusitis induced subdural empyema. Neurology (NY) 1983; 33: 123-32. 12. Carter BL, Bankoff MS, Fisk JD. Computed tomographic detection of sinusitis responsible for intracranial and extracranial infections. Radiology 1983; 147: 739-42. 13. Parker GS, Tami TA, Wilson JF, Fetter TW. Intracranial complications of sinusitis. South Med J 1989; 82: 563-8. 14,. Humphrey MA, Simpson GT, Grindlinger GA. Clinical characteristies of nosocomial sinusitis. Ann Otol Rhinol Laryngol 1987; 96: 687-90. 15. Caplan ES, Hoyt NJ. Nosocomial sinusitis. JAMA 1982; 247: 639-41. 16. Bell RM, Page GV, Bynoe RP, Dunham ME, Brill AH. Posttraumatic sinusitis. J Trauma 1988; 28: 923-30. 17. Bilaniuk LT, Zimmerman RA. Computed tomography in evaluation of the paranasal sinuses. Radiol Clin North Am 1982; 20: 51-66. 18. Vuorinen P, Kauppila A, Pulkkinen K. Comparison of results of roentgen examination and puncture and irrigation of the maxillary sinuses. J Laryngol Otol 1962; 76: 359-64. 19. McNeill RA. Comparison of the findings on transillumination, x-ray and lavage of the maxillary sinus. J Laryngol Otol 1963; 77: 1009-13. 20. Hamory BH, Sande MA, Sydnor A Jr, Scale DL, Gwaltney JM Jr. Etiology and antimicrobial therapy of acute maxillary sinusitis. J Infect Dis 1979; 139: 197-202. 21. Christensen L, Schaffer S, Ross SE. Otitis media in adult trauma patients: incidence and clinical significance. J Trauma 1991; 31: 1543-5. 22. Desmond P, Raman R, Idikula J. Effect of nasogastric tubes on the nose and maxillary sinus. Crit Care Med 1991; 19: 509-11. 23. Stauffer JL, Olson DE, Petty TL. Complications and consequences of endotracheal intubation and tracheostomy. Am J Med 1981; 70: 65-76. 24. Ylikoski J, Savolainen S, Jousimies-Somer H. The bacteriology of acute maxillary sinusitis. Otol Rhinol Laryngol 1989; 51: 175-81. 25. Linden BE, Aguilar EA, Allen SJ. Sinusitis in the nasotracheally intubated patient. Arch Otolaryngol Head Neck Surg 1988; 114: 860-1. 26. Sottile FD, Marrie T J, Prough DS, et al. Nosocomial pulmonary infection: possible etiologic significance of bacterial adhesion to endotracheal tubes. Crit Care Med 1986; 14: 265-70.
DISCUSSION Robert E. Condon (Milwaukee, WI): These researchers found only 26 of 598 intensive care unit (ICU) patients who were potential candidates for the development of sinusitis. The selected subset of 26 patients all had computed tomographic (CT) examinations of the sinuses, and most had pneumatic otoseopy. Fourteen patients had opacification or an air-fluid level in the sinus. Ten positive cultures were found by fluid aspiration. Five other patients had only thickening of the sinus lining noted on CT. A small amount of fluid was aspirated from
THE AMERICAN JOURNAL OF SURGERY VOLUME 164 NOVEMBER 1992 415
BORMAN ET AL
only one of these patients, and, on culture, it was found to be negative. Among this total group of 598 patients, only 1 proved to have fever due to sinusitis alone, which certainly justifies the conclusion expressed in the title. The authors have effectively put to rest one of the many imaginary trolls that lurk under the ICU bridge. The etiology of sinusitis in ICU patients has never be~n obscure. Nasotracheal, nasogastric, and other tubes through the nose are the culprit. Dr. Brown, the role of the various tubes wasn't clear to me. Of the 14 patients with major radiographic signs of sinusitis, what kinds of tubes were in place when the diagnosis was made? Another question I have concerns the recommendation for the use of CT in the diagnosis of sinusitis. I don't agree with the contention that most portable sinus films done at the bedside are of poor quality. Quality is a reflection of the training and the supervision of the radiology technicians. Good technicians produce good sinus films, and they're less expensive and do not require moving the patient out of the ICU. Finally, I want to explore what is apparently your practice in the ICU of keeping nasotracheal and nasogastric tubes in place for relatively long periods of time. Nasotracheal intubation is usually used to establish an airway in patients in whom potential cervical spine injuries exist or in whom an awake induction of anesthesia is needed. These nasotracheal tubes are removed or
416
THE AMERICAN JOURNAL OF SURGERY
exchanged for an oral tube shortly after surgery. Nasotracheal tubes are also needed for patients with jaw and facial injuries and are maintained beyond the initial day in those who have had their mouth wired shut. Most such patients are candidates for early elective tracheostomy. In regard to nasogastric tubes, most ICU patients don't need a nasogastric tube after the second or third day, and many don't need them at all. Nasogastric tubes are often used because we don't consider the alternatives. Phillip M. Brown (closing): Dr. Condon, we agree that transnasal cannulas are undoubtedly a major contributing factor in sinusitis. By virtue of the study design, all patients had at least one transnasal cannula. Patients with nasotracheal tubes had the highest frequency of major radiographic sinusitis. We agree that the need for long-term use of nasotracheal and nasogastric tubes in ICU patients is infrequent. Our experience with bedside portable sinus roentgenograms differs from yours. For whatever reasons, the films are often nondiagnostic. In addition, portable studies do not allow evaluation of the sphenoidal and ethmoidal sinuses, which have frequently been involved in our patients. Since many of these surgical ICU patients with occult fevers were also scheduled for abdominal or chest CT scans, sinus CT was Seldom the sole reason for transporting a patient from ICU to radiology.
VOLUME 164
NOVEMBER 1992
