McCartney, Stephen A. MD, PhD*; Sabo, Michelle C. MD, PhD†; Massad, L. Stewart MD‡; Hagemann, Andrea R. MD‡; Mutch, David G. MD‡; Powell, Matthew A. MD‡; Thaker, Premal H. MD, MS‡; Novetsky, Akiva P. MD, MS‡
Departments of *Obstetrics and Gynecology, and †Internal Medicine, University of Washington, Seattle, WA; and ‡Division of Gynecologic Oncology, Department of Obstetrics and Gynecology, Washington University School of Medicine and Siteman Cancer Center, St Louis, MO.
Address correspondence and reprint requests to Akiva P. Novetsky, MD, MS, Division of Gynecologic Oncology, Department of Obstetrics and Gynecology, Washington University School of Medicine and Siteman Cancer Center, 4911 Barnes Jewish Plaza, Campus Box 8064, St Louis, MO. E-mail address: firstname.lastname@example.org.
Supplemental digital content is available for this article. Direct URL citation appears in the printed text and is provided in the HTML and PDF versions of this article on the journal’s Web site (www.ijgc.net).
The authors declare no conflicts of interest.
Received December 27, 2013
Accepted February 23, 2014
Objectives: The objectives of this study are to identify the characteristics of febrile gynecologic oncology patients and to evaluate the utility of common diagnostic procedures used to assess the etiologies of their fevers.
Methods/Materials: Retrospective data were collected for 200 consecutive patients admitted to the gynecologic oncology service at 1 institution between January 2008 and December 2012 for a diagnosis of fever. Data were collected using contingency tables, and the χ2 test was used as appropriate.
Results: Of the patients admitted for evaluation of fever, 142 (71%) of 200 had a documented fever during hospitalization. The most common etiologies of fever in this population were urinary tract infections (28%) and bloodstream infections (27%), whereas 24% of those admitted for fever did not have a source identified. Abdominal/pelvic computed tomography (CT) scans established the etiology of fever in 53 (60%) of the 89 patients tested, whereas chest x-ray and chest CT were diagnostic for 6% and 21%, respectively. Blood and urine cultures were diagnostic in 29% and 32% of cases, respectively. Patients admitted within 30 days of surgery had a higher percentage of wound infections (38% vs 10%, P < 0.001) as compared with those admitted for more than 30 days after surgery.
Conclusions: The initial evaluation of the febrile gynecologic oncology patient without obvious source by history and examination should include urinalysis with reflex culture and blood cultures. Abdominopelvic and chest CT may be useful when fever persists and initial assessment is unrevealing. Chest x-ray is commonly done but infrequently diagnostic. Wound exploration may be important in patients with fevers for more than 30 days after surgery.
In studies of oncology patients, fever accounted for 12% to 35% of admissions. However, the etiology of fever in this population remains poorly defined.1–3 In addition to infections, fevers have been commonly attributed to medications, neutropenia, and the underlying cancer. Previous studies have demonstrated that blood cultures are diagnostic in 25% of cases, whereas x-rays yield diagnostic information in only 3% of cases.4–6 These studies may not be generalizable to gynecologic oncology patients, among whom the precise etiologies of fever have not been rigorously studied. One report found that more than half of febrile gynecology patients had no identifiable source.7 There are limited data on the utility of commonly used laboratory and imaging studies. Because an improved understanding of fever may help direct diagnostic testing, improve patient care, and reduce health care expenditures, we aimed to identify the etiologies of fevers and examine the utility of diagnostic testing for a cohort of patients admitted for fever to the gynecologic oncology service of a tertiary care medical center.
MATERIALS AND METHODS
Study Participants and Clinical Data
This retrospective study evaluated 200 consecutive patients admitted to the gynecologic oncology service at Barnes-Jewish Hospital between January 2008 and December 2012 for evaluation of fever. After the approval of the study by the Washington University School of Medicine institutional review board, patients were identified using admission and discharge logs. Inclusion criteria included the following: known diagnosis of gynecologic cancer, chief complaint or report of subjective or objective fever in the admission history, or documented fever on admission. Exclusion criteria included fever during an admission for surgery or hospital discharge after less than 24 hours of observation. Demographic and medical data were abstracted from the electronic medical record. Temperature was measured orally, and fever was defined as any temperature greater than or equal to 38°C. Mild neutropenia was defined as neutrophil count less than 1500 cells/µL, and severe neutropenia was defined as neutrophil count less than 500 cells/µL.
Upon admission, all patients underwent a thorough history and physical examination. The following results on laboratory or imaging tests were considered diagnostic: (1) chest x-rays where the report identified pneumonia or other infection; (2) computed tomography (CT) scans of the abdomen/pelvis with an impression on the report suggesting infection, including abscess or fat stranding; (3) transthoracic and/or transesophageal echocardiography were considered positive if there were findings of valvular vegetation and were considered diagnostic for endocarditis with positive findings in the setting of positive blood cultures; (4) C. difficile toxin assay that was positive; (5) urinalyses with leukocyte esterase greater than or equal to 2 or the presence of nitrites were considered positive and were considered diagnostic in the presence of a positive urinary culture; and (6) blood, urinary, and wound cultures were considered positive if the presence of any organism was detected above the laboratory threshold and were considered diagnostic based on the pattern of growth and organisms grown, in the context of all clinical factors.
Statistical analysis data were collected and analyzed using contingency tables. Pearson χ2 and Fisher exact tests were used to compare dichotomous variables. P < 0.05 was considered statistically significant. Statistical analyses were performed using Stata v9.0 (College Station, TX).
Patient characteristics at the time of admission for fever are summarized in Tables 1 and 2. The average patient age at presentation was 57 years. Only 51 patients (26%) had fever documented at admission; however, 142 patients (71%) had a documented fever during their hospital stay. Prior recent admission was common, with 118 (59%) of patients previously admitted within 30 days of the index admission. Of the 166 patients with prior surgery, only 29 (15%) were admitted within 30 days of a prior operation. A total of 102 women (51%) had received chemotherapy within 30 days of admission, and 73 (72%) of these were admitted within 12 days of chemotherapy. The most common day of admission postchemotherapy was 8 days (Fig. 1, Supplemental Digital Content 1, http://links.lww.com/IGC/A216). Mild and severe neutropenia were documented in 43 (22%) and 25 (13%) of patients, respectively.
The etiologies of fever are summarized in Table 3. The most common diagnoses were infections of the urinary tract and bloodstream, comprising 28% and 27% of all patients, respectively. Skin and wound infections were found in 13% of patients, and pneumonia was found in only 5% of all patients. Iatrogenic causes of fever, such as drug allergy and transfusion reaction, were uncommon. Sixty-six patients had more than 1 diagnosis, most commonly a urinary tract infection (UTI) in addition to another diagnosis.
The etiology of fever was not determined in 48 patients (24%). Among these patients, 16 were readmitted within 30 days with recurrent fevers, and in 7 (44%) of these patients, a source of fever was detected upon readmission, leaving only 41 (21%) with no diagnosis after evaluation.
The etiologies of fever by primary tumor site are summarized in Table 3 and supplemental Table 1 (Supplemental Digital Content 1, http://links.lww.com/IGC/A216). Patients with ovarian cancer had a significantly higher rate of pneumonia as compared with patients with cervix and uterine cancer (11% vs 2% and 0%, respectively; P = 0.007). Patients with cervical and uterine cancer were more likely than those with ovarian cancer to be diagnosed with infections of the urinary tract (42% and 29% vs 13%, respectively; P = 0.001). No other differences in the etiology of fevers were seen among the different primary cancers.
We assessed whether the etiology of fever differed based on common patient characteristics (Table 4). Patients with neutropenia had a higher rate of unknown diagnosis than the overall study population (42% vs 19%, P = 0.002). Compared with patients admitted for more than 30 days after surgery, those admitted within 30 days of surgery had a higher percentage of wound infections (38% vs 10%, P < 0.001). There was no difference in the incidence of abscess, blood stream infection, or pneumonia in patients who had previously received chemotherapy or those admitted with neutropenic fever as compared with the rest of the cohort.
Diagnostic studies are summarized in Table 5. Blood cultures, urine cultures, or wound/abscess cultures were performed in 94%, 85%, and 26% of patients, respectively. Of the urine cultures, 56 (32.7%) were positive, and 55 (32.1%) were diagnostic. Urinalysis result was positive in 53% of patients tested; however, only 30% were found to have positive urine cultures. The incidence of a positive urine culture in the setting of a negative urinalysis was 8%. Blood cultures were positive in 63 patients (33.7%) and diagnostic in 54 patients (28.9%). The most frequently isolated organisms from blood cultures were coagulase-negative staphylococci and Staphylococcus aureus (Fig. 1). The most common Gram-negative organisms were Klebsiella and E. coli, although there were fewer overall infections with these organisms.
Most patients also underwent imaging studies, with 153 patients (77%) receiving a chest x-ray and 107 (54%) receiving at least 1 CT study (chest, abdomen/pelvis, or head/neck). Nine chest x-rays met the criteria for “diagnostic” (6%) and contributed to a final diagnosis of pneumonia in 5% of women. The diagnostic yield for chest CT was 21%. Of the 89 abdomen/pelvis CT performed, 53 (60%) were considered diagnostic. Head and neck CT was diagnostic in 1 (6%) of 18 instances.
We also assessed the utility of diagnostic testing in patient subgroups (Tables 1–3, Supplemental Digital Content 1, http://links.lww.com/IGC/A216). There were no additional statistically significant differences in the performance of diagnostic testing in neutropenic patients or in patients presenting 30 days after surgery or chemotherapy.
Initial management of gynecologic oncology patients admitted with fever should include blood and urine cultures before administration of antibiotics that cover the broad spectrum of organisms we identified, including Gram-positive and Gram-negative pathogens. If these steps are nondiagnostic and fever persists, abdominopelvic CT scanning will reveal the etiology of fever in many women. This approach seems valid across compared patient groups, including neutropenic, postsurgical, and postchemotherapy patients, except that women presenting within 30 days of surgery should have wounds explored.
Compared with other studies that focused strictly on nongynecologic patients with neutropenic or postoperative fever, this study included all gynecologic oncology patients admitted with a diagnosis of fever. Overall, fever etiologies were similar between the different gynecologic malignancies, with 2 notable exceptions. Ovarian cancer patients had a higher incidence of pneumonia, whereas uterine and cervical cancer patients were found to have a higher incidence of UTIs. These differences may be related to the systemic spread of ovarian cancer with pleural effusions leading to an increased risk for pneumonia as compared with more localized disease in uterine and cervical cancers. Adjuvant treatment with radiation therapy could also result in increased risk for infections of the urinary tract.
In our population, an etiology for fever was determined for 80% of patients. This proportion is greater than that in previous studies of febrile neutropenic patients, in which the etiology of fever remained unknown after evaluation in more than 50% of patients.6–9 However, we also found that neutropenic patients often had no identifiable source of fever, suggesting that fevers may arise from compromised mucosal immunity in these women. Blood cultures were positive in 34% and diagnostic in 29% of our patients. The utility of blood cultures in this study is higher than what has previously been seen in oncology patients8,10 and may reflect the increasing use of central venous catheters in gynecologic oncology patients
Urinalysis has been shown to have a sensitivity of 68% to 98% and specificity of 59% to 96% in various studies.11–13 However, urinalyses in hospitalized patients have lower reliability, and urine culture is recommended for definitive diagnosis in these patients.13 Our finding that only 60% of women with positive urinalysis results had positive urine cultures may reflect contamination from vaginal sources or administration of antibiotics before culture.
Severely neutropenic women made up 13% of all gynecologic oncology patients admitted with fever. Febrile neutropenic gynecologic oncology patients with fever also had a lower percentage of bloodstream and UTIs compared with the overall population. For these women, broad-spectrum antibiotic coverage until resolution of neutropenia may be sufficient. Although prospective studies are needed, hospitalization may not be required for many febrile neutropenic gynecologic oncology patients.14,15
Although 60% of patients undergoing abdominal/pelvic CT scans had findings consistent with infection, several of these patients already had confirmed infection from other sources. In addition, not all of these CT scans were performed for diagnosis of fever because several patients had coexisting small bowel or urinary tract obstruction; therefore, the true rate of diagnosis attributable to CT scans is unclear in retrospect. Chest CT had an overall lower rate of utility than abdominal/pelvic CT in this population. Head/neck CT had a very poor yield in the evaluation of fever. Prospective assessment of protocols incorporating CT imaging may better define the role of these tests in febrile gynecologic oncology patients.
This study has a number of limitations. In addition to its retrospective nature, this study assessed a heterogeneous population of 200 consecutive patients admitted for fever, including those who were recently postoperative, those receiving chemotherapy, or those with febrile neutropenia. The small numbers of women in subgroups limit our ability to perform subgroup analyses. Multicenter studies focused on various classes of febrile gynecologic oncology patients may be needed to define optimal management protocols. Finally, the presence of central lines and indwelling vascular catheters was poorly documented in our medical records precluding and assessment of these as a source of an infectious fever.
Despite these limitations, our study defines the characteristics of febrile gynecologic oncology patients and provides an evidence-based approach to stepwise assessment and management of these women. Research is ongoing to determine whether some gynecologic oncology patients with fever can be evaluated as outpatients.
1. Aprile G, Pisa FE, Follador A, et al. Unplanned presentations of cancer outpatients: a retrospective cohort study. Support Care Cancer. 2013; 21: 397–404.
2. Majem M, Galan M, Perez FJ, et al. The oncology acute toxicity unit (OATU): an outpatient facility for improving the management of chemotherapy toxicity. Clin Transl Oncol. 2007; 9: 784–788.
3. McKenzie H, Hayes L, White K, et al. Chemotherapy outpatients’ unplanned presentations to hospital: a retrospective study. Support Care Cancer. 2011; 19: 963–969.
4. Feusner J, Cohen R, O’Leary M, et al. Use of routine chest radiography in the evaluation of fever in neutropenic pediatric oncology patients. J Clin Oncol. 1988; 6: 1699–1702.
5. Oude Nijhuis CS, Gietema JA, Vellenga E, et al. Routine radiography does not have a role in the diagnostic evaluation of ambulatory adult febrile neutropenic cancer patients. Eur J Cancer. 2003; 39: 2495–2498.
6. Hakim H, Flynn PM, Knapp KM, et al. Etiology and clinical course of febrile neutropenia in children with cancer. J Pediatr Hematol Oncol. 2009; 31: 623–629.
7. McMeekin DS, Gazzaniga C, Berman M, et al. Retrospective review of gynecologic oncology patients with therapy-induced neutropenic fever. Gynecol Oncol. 1996; 62: 247–253.
8. Freifeld AG, Bow EJ, Sepkowitz KA, et al. Clinical practice guideline for the use of antimicrobial agents in neutropenic patients with cancer: 2010 update by the Infectious Diseases Society of America. Clin Infect Dis. 2011; 52: e56–e93.
9. Pizzo PA. Management of fever in patients with cancer and treatment-induced neutropenia. N Engl J Med. 1993; 328: 1323–1332.
10. Bates DW, Cook EF, Goldman L, et al. Predicting bacteremia in hospitalized patients. A prospectively validated model. Ann Intern Med. 1990; 113: 495–500.
11. Bent S, Nallamothu BK, Simel DL, et al. Does this woman have an acute uncomplicated urinary tract infection? JAMA. 2002; 287: 2701–2710.
12. Pappas PG. Laboratory in the diagnosis and management of urinary tract infections. Med Clin North Am. 1991; 75: 313–325.
13. Wilson ML, Gaido L. Laboratory diagnosis of urinary tract infections in adult patients. Clin Infect Dis. 2004; 38: 1150–1158.
14. Klastersky J, Paesmans M, Georgala A, et al. Outpatient oral antibiotics for febrile neutropenic cancer patients using a score predictive for complications. J Clin Oncol. 2006; 24: 4129–4134.
15. Innes H, Lim SL, Hall A, et al. Management of febrile neutropenia in solid tumours and lymphomas using the Multinational Association for Supportive Care in Cancer (MASCC) risk index: feasibility and safety in routine clinical practice. Support Care Cancer. 2008; 16: 485–491.
© 2014 by the International Gynecologic Cancer Society and the European Society of Gynaecological Oncology.