Abstract
Objectives
Febrile neutropenia is considered an oncologic emergency, for which prompt initiation of antibiotics is essential.
Methods
We conducted a retrospective cohort study for the 2006 calendar year involving all adult oncology patients presenting with febrile neutropenia to a regional health authority’s emergency departments. The objective was to determine the time from triage to antibiotic administration and its impact on patient outcomes.
Results
We identified 68 patients presenting with febrile neutropenia, most of whom (76%) were seen in tertiary care centers. Of those patients, 65% were triaged to be seen within 15 minutes of arrival in the emergency room; however, the median time to reassessment was 57 minutes. The median time from triage to antibiotic administration was 5 hours (range: 1.23–22.8 hours). No increased risk of death or increased length of hospital stay was associated with delayed antibiotic administration. Older patients and patients without caregiver support were more likely to experience delayed antibiotic administration (odds ratio: 3.8 and 12.7 respectively).
Conclusions
We were not able to show a deleterious effect of delay in antibiotic administration, but our analysis identified several points at which patient flow through the emergency room could be improved.
Keywords: Antineoplastic chemotherapy, febrile neutropenia, emergency department
1. INTRODUCTION
Neutropenia is one of the major dose-limiting toxicities of systemic cancer chemotherapy. The depth and duration of neutropenia correlate with the risk of developing infection and death 1.
Episodes of febrile neutropenia (fn) are considered an oncologic emergency. Among cancer patients presenting with an episode of fn, one series found that 30% had microbiologically confirmed bacteremia 2. Another group identified a 10% prevalence of gram-negative bacterial infections 3. An episode of fn is a potentially life-threatening condition 4, and therefore prompt evaluation and timely antibiotic administration are considered crucial.
Several studies have described wait times for adult oncology patients presenting to emergency departments (eds) with episodes of fn. One group identified a 75-minute median wait time from ed admission to examination, with the median time to antibiotic administration being 210 minutes. Patients with advanced cancer and those with more comorbidities waited longer than did those with earlier-stage, less complex disease 5. Another group identified a 107-minute median wait time from triage to antibiotic administration for adult patients presenting to the ed with an episode of fn 6. A prospective study examined the time from presentation in the ed to antibiotic administration in patients with a presumptive serious infection in need of immediate empiric antibiotic therapy. The median time to antibiotic administration was 5 hours 7. In the setting of sepsis and hypotension, a delay in antibiotic administration is associated with a survival decrement of 7.5% per hour of delay 8.
Currently, a limited number of benchmarks have been established for evaluation of adult patients with fn in Canada, but several Canadian pediatric centres have developed benchmarks for the evaluation and therapy of pediatric patients with fn. The Child Health Program of Manitoba benchmarks for all pediatric patients presenting with the diagnosis of fn to the ed are these:
To be assessed within 15 minutes by a nurse and a physician [Canadian Triage and Acuity Scale (ctas) level 2 9]
To commence antibiotic administration within 2 hours of initial presentation 10
Similar benchmarks are in use in other centres across Canada (Chagla Y. Canadian Nosocomial Surveillance Program. Personal communication).
In the current study, we retrospectively analyzed the initial management of all adult oncology patients presenting to eds within a regional health authority over the course of 1 year.
2. METHODS
2.1 Cohort Assembly
A population at risk was assembled by identifying all adult patients (18 years of age and over) within the Manitoba cancer registry who received systemic chemotherapy between November 1, 2005, and December 31, 2006. Each hospital within the Winnipeg Regional Health Authority (wrha) generated a list of all adult admissions for the management of neutropenia (ICD10 code D70) between January 1 and December 31, 2006. Patients were included if they had an absolute neutrophil count less than 1.5×109/L with a temperature of at least 38.0°C at the time of triage or within the 24 hours preceding their presentation to the ed. Cases were excluded if the episode of fn developed in hospital rather than in the ed, if fever was absent, or if the admission for fn occurred directly from clinic to a hospital ward. The final cohort consisted of the adult patients who had received systemic chemotherapy within the 60 days preceding presentation to the ed for an episode of fn. A professional data abstractor then undertook a detailed chart review for those patients. Triage assignment in the ed was determined by the ctas score (1, immediate assessment; 2, within 15 minutes; 3, within 30 minutes; 4, within 1 hour) 8. Vital status was ascertained at 5 and 30 days after the ed visit, through linkage with the provincial cancer registry. Ethics approval for the study was granted by the Health Research Ethics Board of the University of Manitoba and by the Research Review Committee of the wrha.
2.2 Analysis
Two groups were identified: patients receiving timely antibiotics (less than 5 hours from ed triage to antibiotic administration), and those not receiving timely antibiotics. Univariate logistic regression was used to compare the groups. Variables on univariate analysis with a p value less than 0.1 were then entered into a multivariate logistic analysis. Multivariate logistic regression was performed to identify predictors of delayed antibiotic administration. Univariate logistic regression was used to examine whether delayed antibiotic administration affected vital status at 5 and 30 days after ed triage. A Kaplan–Meier survival analysis 11 was used to determine whether time to antibiotic administration was different for patients with hematologic and nonhematologic malignancies.
All statistical analyses were performed using the SAS software application (version 9.1: SAS Institute, Cary, NC, U.S.A.). Unless otherwise indicated, data are presented as medians or means with ranges.
3. RESULTS
3.1 Baseline Cohort Description
We identified 68 oncology patients who, between January 1 and December 31, 2006, experienced an episode of fn and underwent assessment through an ed within a wrha facility. Neutrophil counts for the cohort were distributed as follows:
39 patients (57%) presented to the ed with an absolute neutrophil count below 0.5×109/L.
23 patients (34%) had an initial neutrophil count between 0.5 and 1.0×109L.
6 patients (9%) had only an initial white blood cell count available, which ranged from 1.0 to 1.2×109L.
Of the 68 patients, 6 (9%) visited the ed more than once in close proximity to the fn visit for which antibiotics were administered. In the cohort, 27 patients (40%) had a hematologic malignancy [the most common diagnosis being large cell lymphoma (8 patients)] and 41 patients (60%) had a solid malignancy [the most common diagnosis being breast cancer (27 patients, Table I)]. The median age in the cohort was 59.5 years (range: 20–83 years), and 30 of 67 patients (45%) were men. Most members of the cohort (52 patients, 76%) were seen in a tertiary care facility. Filgrastim was administered in 42 patients (62%) before their ed presentation, and filgrastim was started in 6 patients (9%) after ed presentation. A central venous catheter was present in situ in 46 patients (68%). Median duration from last chemotherapy to ed triage was 8.5 days (range: 1–42 days). Almost two thirds of the cohort (44 patients, 65%) were assigned a ctas score of 2 (see Table I). After ed triage and assessment, 2 patients (3%) with fn required admission to an intensive care unit. In 15 patients (22%), blood cultures were positive, with 9 patients (13%) having gram-negative bacterial isolates.
TABLE I.
Baseline characteristics
| Variable | Value |
|---|---|
| Patients (n) | 68 |
| Sex (n) | |
| Men | 30 |
| Women | 38 |
| Age (years) | |
| Median | 59.5 |
| Range | 20–83 |
| Disease (n) | |
| Hematologic malignancy | 27 |
| Multiple myeloma | 6 |
| Large cell lymphoma | 9 |
| High-grade lymphoma/all | 3 |
| Other | 9 |
| Solid tumors | 41 |
| Breast cancer | 27 |
| gi malignancy | 8 |
| Other | 6 |
| cvcin situ [n (%)] | 46/68 (68) |
| Days from last chemotherapy | |
| Median | 8.5 |
| Range | 1–42 |
| ctas score = 2 [n/N (%)] | 44/68 (65) |
| Absolute neutrophil count | |
| Median | 0.23 |
| Range | 0–1.5 |
| Triage mean blood pressure | |
| Median | 84 |
| Range | 49–118 |
| Maximal ed temperature (°C) | |
| Median | 38.7 |
| Range | 38–40 |
| Time (minutes) from ed triage to Nurse assessment | |
| Median | 57 |
| Range | 2–335 |
| emo assessment | |
| Median | 60 |
| Range | 7–427 |
| Antibiotic administration | |
| Median | 616 |
| Range | 74–1369 |
| Hospital stay (days) | |
| Median | 5.1 |
| Range | 0.2–55.4 |
all = acute lymphocytic leukemia; gi = gastrointestinal; cvc = central venous catheter; ctas = Canadian Triage and Acuity Scale; ed = emergency department; emo = emergency medical officer.
3.2 Time to Antibiotic Administration
The median time from ed triage to antibiotic administration was just over 5 hours (range: 1.23–22.8 hours). In our cohort, only 4 patients (6%) were administered antibiotics within 2 hours of triage. A log rank test showed that the median times to antibiotic administration of 6 hours for those with hematologic malignancies and 5 hours for those with nonhematologic malignancies (by Kaplan–Meier time-to-event analysis) were not significantly different.
To examine factors predictive for delayed antibiotic administration, we divided the cohort into those receiving antibiotics within 5 hours of triage and those receiving antibiotics more than 5 hours after triage.
3.2.1 Predictors of Delayed Antibiotic Administration
On both univariate and multivariate analysis, age greater than 60 years and lack of a caregiver (either a family member or a friend) were associated with delayed antibiotic administration, with odds ratios of 3.8 (95% confidence interval: 1.3 to 11) and 12.7 (95% confidence interval: 1.4 to 112) respectively (Tables II and III).
TABLE II.
Univariate logistic regression examining predictors of delayed antibiotic administrationa
| Variable | or | 95% ci | p Value |
|---|---|---|---|
| Sex (men vs. women) | 0.85 | 0.3 to 2.2 | 0.74 |
| Age (<60 vs. ≥60 years) | 3.1 | 1.1 to 8.6 | 0.03 |
| Diagnosis (hematologic vs. solid) | 0.7 | 0.3 to 1.9 | 0.51 |
| Caregiver status (family- vs. self-care) | 9.6 | 1.1 to 81 | 0.04 |
| Prior ed notification (yes vs. no) | 2.2 | 0.6 to 8.2 | 0.21 |
| Patient history of fn episodes (yes vs. no) | 1.0 | 0.4 to 3.0 | 0.94 |
| Central venous catheter (present vs. absent) | 0.9 | 0.29 to 2.8 | 0.86 |
| ctas score (2 vs. ≥2) | 2.2 | 0.82 to 6.0 | 0.12 |
| ed maximal temperature (<39°C vs. ≥39°C) | 0.5 | 0.19 to 1.3 | 0.17 |
| Triage mean bp (<90 vs. ≥90 mmHg) | 1.2 | 0.41 to 3.3 | 0.76 |
| Triage heart rate | 0.99 | 0.97 to 1.0 | 0.55 |
| Triage respiratory rate | 0.91 | 0.79 to 1.0 | 0.17 |
| Absolute neutrophil count | 0.67 | 0.23 to 1.9 | 0.45 |
More than 5 hours after emergency department triage. or = odds ratio; ci = confidence interval; ed = emergency department; fn = febrile neutropenia; ctas = Canadian Triage and Acuity Scale; bp = blood pressure.
TABLE III.
Multivariate logistic regression examining predictors of delayed antiobiotic administrationa
| Variable | or | 95% ci |
|---|---|---|
| Age (≥60 vs. <60 years) | 3.8 | 1.3 to 11 |
| Caregiver status (self vs. family) | 12.7 | 1.4 to 112 |
More than 5 hours after emergency department triage. or = odds ratio; ci = confidence interval.
3.2.2 Time to Antibiotic Administration and Patient Outcomes
The proportion of patients alive at 5 days and 30 days after ed triage were 0.94 and 0.91 respectively. In univariate logistic regression, we observed no increased risk of death associated with delayed antibiotic administration. The median length of hospital stay for the entire cohort was 5 days (range: 0.2–55 days). Delayed antibiotic administration did not lead to an increased length of hospital stay.
4. DISCUSSION
During the period studied, the median time to antibiotic administration was 5 hours in adult oncology patients with an episode of fn. Only 4 of 68 patients received antibiotics within 2 hours of ed triage. Most patients (44 of 68, 65%) were triaged in the ed with a ctas score of 2 (suggesting reassessment within 15 minutes). However, for that group, the median wait to be reassessed by a nurse was 57 minutes. The median wait for a laboratory evaluation, including blood draw for culture, was 101 minutes. The primary oncologist or the oncology team notified the ed of the patient’s impending arrival in 11 cases (16%).
The Infectious Diseases Society of America guideline on the management of the neutropenic cancer patient 12 and the wrha Guidelines for the Management and Monitoring of Low-Risk versus High-Risk Febrile Neutropenic Cancer 13 suggest prompt initiation of antibiotics after a timely evaluation. Several pediatric ed programs in Canada have adopted standardized protocols to ensure timely antibiotic administration in pediatric patients with fn. Within the regional health authority institutions in which our study was performed, the pediatric ed and oncology programs introduced a 2-hour benchmark for the time from triage to antibiotic administration. At the same time, a clinical pathway was established, with standard preprinted order sets and an algorithm. Before introduction of the benchmark, median time to antibiotic administration was 5 hours; after adoption of the benchmark, 90% of pediatric patients received antibiotics within 2 hours of ed presentation 10.
The delay in antibiotic administration noted in the present study is likely multifactorial; however, development of a standardized protocol to direct the flow of patients with an episode of fn through the ed would be of significant benefit. The adoption of algorithms, protocols, or standardized order sets within a variety of institutional settings has clearly been shown to lower the time to antibiotic administration in the setting of an episode of fn 6,10. Within the broader scope of emergency care, several studies have identified the roles that quality improvement projects, guideline development, and clinical pathway implementation play in improving outcomes in the management of acute myocardial infarction and cerebrovascular disease, among other conditions 14–16.
Patients who were older and who did not have caregiver support were more likely to wait longer from triage to antibiotic administration. However, delayed antibiotic administration did not translate into decreased survival or increased length of hospital stay. Oncology patients with fn are on a spectrum of acuity, with some able to manage as outpatients, and others presenting in septic shock. In our study cohort, 13% of patients had cultures positive for gram-negative bacteria. Because patients with fn may worsen acutely, urgent assessment and timely antibiotic administration is justified.
We could identify no patients in the study cohort who were undergoing high-dose cytarabine consolidation in the management of acute myeloid leukemia. Consolidation therapy with high-dose cytarabine is associated with a 60% risk of fn 17,18. Patients with acute myeloid leukemia undergoing consolidation receive outpatient treatment in Manitoba, in a process that is highly protocolized and that follows a care map. That patient group is monitored daily from initiation of chemotherapy until count recovery. They are treated prophylactically with antibiotics and growth factors, and with onset of an episode of fn, they are rapidly admitted directly to a specialized ward within the wrha. Within the context of the wrha, discussion is ongoing concerning implementation of a care map and preprinted order sets to complement the existing regional clinical practice guideline 13 so that the management of oncology patients presenting to the ed with fn is optimized.
Our study has several limitations. Because of its retrospective nature, we could not accurately capture patient comorbidity or performance status, both of which affect outcomes in patients with fn 19. We defined timely antibiotic administration as being within 5 hours from triage. The literature holds no definitive guidelines concerning the optimal time to administer antibiotics in the setting of an episode of fn. Various experts have used the 2-hour benchmark from time of triage to antibiotic administration as a treatment target. Within our study cohort, only 6% received antibiotics within 2 hours of ed triage. The 5-hour time point represents the median time to antibiotic administration in this cohort of 68 patients, and it is thus a pragmatic cut-off for our analysis. The small number of patients in the cohort limits the study’s power to demonstrate a detrimental effect of delayed antibiotic administration.
We were not able to demonstrate an adverse effect of delayed antibiotic administration in oncology patients with an episode of fn. Nevertheless, patient care could be improved by implementing various strategies to shorten the time from triage to treatment. Protocolizing ed care and producing a care map for these vulnerable patients would streamline their care and provide a framework for the ed staff in managing an uncommon ed event.
5. ACKNOWLEDGMENTS
Data assembly for this project was undertaken by the Audits and Quality analysis group of the Winnipeg Regional Health Authority (G. Taylor, E. McIntyre, C. Rogowski). The authors thank E.F. Cook for careful reading of the manuscript.
Footnotes
6. CONFLICT OF INTEREST DISCLOSURES
The authors have no financial conflicts of interest to declare.
7. REFERENCES
- 1.Bodey GP, Buckley M, Sathe YS, Freireich EJ. Quantitative relationships between circulating leukocytes and infection in patients with acute leukemia. Ann Intern Med. 1966;64:328–40. doi: 10.7326/0003-4819-64-2-328. [DOI] [PubMed] [Google Scholar]
- 2.Viscoli C, Bruzzi P, Castagnola E, et al. Factors associated with bacteraemia in febrile, granulocytopenic cancer patients. The International Antimicrobial Therapy Cooperative Group (IATCG) of the European Organization for Research and Treatment of Cancer (EORTC) Eur J Cancer. 1994;30A:430–7. doi: 10.1016/0959-8049(94)90412-X. [DOI] [PubMed] [Google Scholar]
- 3.Cordonnier C, Herbrecht R, Buzyn A, et al. Risk factors for gram-negative bacterial infections in febrile neutropenia. Haematologica. 2005;90:1102–9. [PubMed] [Google Scholar]
- 4.Klastersky J. Management of fever in neutropenic patients with different risks of complications. Clin Infect Dis. 2004;39(suppl 1):S32–7. doi: 10.1086/383050. [DOI] [PubMed] [Google Scholar]
- 5.Nirenberg A, Mulhearn L, Lin S, Larson E. Emergency department waiting times for patients with cancer with febrile neutropenia: a pilot study. Oncol Nurs Forum. 2004;31:711–15. doi: 10.1188/04.ONF.711-715. [DOI] [PubMed] [Google Scholar]
- 6.Baltic T, Schlosser E, Bedell MK. Neutropenic fever: one institution’s quality improvement project to decrease time from patient arrival to initiation of antibiotic therapy. Clin J Oncol Nurs. 2002;6:337–40. doi: 10.1188/02.CJON.337-340. [DOI] [PubMed] [Google Scholar]
- 7.Natsch S, Kullberg BJ, van der Meer JW, Meis JF. Delay in administering the first dose of antibiotics in patients admitted to hospital with serious infections. Eur J Clin Microbiol Infect Dis. 1998;17:681–4. doi: 10.1007/s100960050160. [DOI] [PubMed] [Google Scholar]
- 8.Kumar A, Roberts D, Wood KE, et al. Duration of hypotension before initiation of effective antimicrobial therapy is the critical determinant of survival in human septic shock. Crit Care Med. 2006;34:1589–96. doi: 10.1097/01.CCM.0000217961.75225.E9. [DOI] [PubMed] [Google Scholar]
- 9.Beveridge R. CAEP issues. The Canadian Triage and Acuity Scale: a new and critical element in health care reform. Canadian Association of Emergency Physicians. J Emerg Med. 1998;16:507–11. doi: 10.1016/s0736-4679(98)00031-6. [DOI] [PubMed] [Google Scholar]
- 10.Cronin G. Balanced Score Card—Child Health Program Winnipeg Regional Health Authority. Winnipeg, MB: Winnipeg Regional Health Authority; September 30. 2006. [Google Scholar]
- 11.Kaplan EL, Meier P. Nonparametric estimation from incomplete observations. J Am Stat Assoc. 1958;53:457–81. doi: 10.2307/2281868. [DOI] [Google Scholar]
- 12.Hughes WT, Armstrong D, Bodey GP, et al. 2002 Guidelines for the use of antimicrobial agents in neutropenic patients with cancer. Clin Infect Dis. 2002;34:730–51. doi: 10.1086/339215. [DOI] [PubMed] [Google Scholar]
- 13.Bow EJ, Ramesar K. Guidelines for the Management and Monitoring of Low-Risk Versus High-Risk Febrile Neutropenic Cancer Patients Managed at Cancer Care Manitoba, the Health Sciences Centre, and St. Boniface General Hospital. Winnipeg, MB: Winnipeg Regional Health Authority; 2005. [Google Scholar]
- 14.Bradley EH, Herrin J, Elbel B, et al. Hospital quality for acute myocardial infarction: correlation among process measures and relationship with short-term mortality. JAMA. 2006;296:72–8. doi: 10.1001/jama.296.1.72. [DOI] [PubMed] [Google Scholar]
- 15.Wang TY, Fonarow GC, Hernandez AF, et al. The dissociation between door-to-balloon time improvement and improvements in other acute myocardial infarction care processes and patient outcomes. Arch Intern Med. 2009;169:1411–19. doi: 10.1001/archinternmed.2009.223. [DOI] [PubMed] [Google Scholar]
- 16.LaBresh KA, Reeves MJ, Frankel MR, Albright D, Schwamm LH. Hospital treatment of patients with ischemic stroke or transient ischemic attack using the “Get With the Guidelines” program. Arch Intern Med. 2008;168:411–17. doi: 10.1001/archinternmed.2007.101. [DOI] [PubMed] [Google Scholar]
- 17.Mayer RJ, Davis RB, Schiffer CA, et al. Intensive postremission chemotherapy in adults with acute myeloid leukemia. Cancer and Leukemia Group B. N Engl J Med. 1994;331:896–903. doi: 10.1056/NEJM199410063311402. [DOI] [PubMed] [Google Scholar]
- 18.Sierra J, Szer J, Kassis J, et al. A single dose of pegfilgrastim compared with daily filgrastim for supporting neutrophil recovery in patients treated for low-to-intermediate risk acute myeloid leukemia: results from a randomized, double-blind, phase 2 trial. BMC Cancer. 2008;8:195. doi: 10.1186/1471-2407-8-195. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Klastersky J, Paesmans M, Rubenstein EB, et al. The Multinational Association for Supportive Care in Cancer risk index: a multinational scoring system for identifying low-risk febrile neutropenic cancer patients. J Clin Oncol. 2000;18:3038–51. doi: 10.1200/JCO.2000.18.16.3038. [DOI] [PubMed] [Google Scholar]