Abstract
Adequate bone marrow recovery is a discharge requirement after admission for febrile neutropenia in oncology patients, without specific threshold in consensus guidelines. In January 2016, our institution implemented count recovery criteria of absolute neutrophil count (ANC) ≥ 100 cells/μL and absolute phagocyte count (APC) ≥ 300 cells/μL compared to prior criteria of ANC ≥ 500 cells/μL. Retrospective analysis comparing pre (July 2013-December 2015, N=68) and post (January 2016- June 2018, N=30) groups showed no difference in readmissions (P>0.9), no patient deaths, and decreased average length of stay in the post group (P<0.0001). Updated count recovery criteria appear feasible and safe.
Keywords: fever and neutropenia, pediatric oncology, discharge
Introduction
Fever and neutropenia (FN) is a common treatment-related complication in pediatric oncology. While the Infectious Diseases Society of America practice guidelines have focused on adult patients, the International Pediatric Fever and Neutropenia Guideline Panel is a multidisciplinary and multinational group that provided updated pediatric specific FN guidelines in 2017.1,2,3 Standard of care typically requires admission for intravenous (IV) antibiotics until resolution of fever and evidence of bone marrow recovery.3,4,5,6 Recommendations include that FN patients should be afebrile ≥ 24 hours with negative blood cultures ≥ 48 hours before discharge.3 However, there is no criterion for adequate bone marrow recovery in consensus guidelines.3,4,5,6
Absolute neutrophil count (ANC) is a marker of adequate bone marrow recovery, and patients with ANC ≤ 100 cells/microliter (μL) have the highest infections risks.6 Recent data suggest ANC of ≥ 100 cells/μL may be an adequate marker of bone marrow recovery for discharge in pediatric oncology FN patients.4,7 Other markers may indicate bone marrow recovery prior to rise in ANC, such as the absolute phagocyte count (APC).8 Incorporation of the APC may provide a more robust picture of count recovery.
In January 2016, the Medical University of South Carolina (MUSC) Division of Pediatric Hematology and Oncology implemented new discharge criteria for FN patients. Criteria continued to include a minimum 24-hour afebrile period and negative blood culture for 48 hours, but adequate bone marrow recovery was defined as ANC ≥ 100 cells/μL and APC ≥ 300 cells/μL. Prior to January 2016, adequate bone marrow recovery was defined as ANC ≥ 500 cells/μL.
2. Methods
Institutional Review Board approval was obtained for retrospective chart review. Group 1 (pre) had defined markers of bone marrow recovery as ANC ≥ 500 cells/μL, with data collection from July 2013 – December 2015. Group 2 (post) had defined markers of bone marrow recovery as ANC ≥ 100 cells/μL and APC ≥ 300 cells/μL, with data collection from January 2016 – June 2018. ANC was calculated via laboratory automation, and APC was calculated by hand with the following equation: APC in cells/μL = [total white blood cell count (103 cells/μL) x (% neutrophils + % bands + % monocytes)]. For example, if a total white blood cell count reported is 1.5 x 103 cells/μL, with neutrophil percentage 5.5%, band percentage 1.4%, and monocyte percentage 12.0%, the APC = [(1.5 x 103 cells/μL) x (0.055 + 0.014 + 0.12)] = 283.5 cells/μL. FN was defined as a single reported or documented temperature ≥ 38 degrees Celsius and ANC ≤ 500 cells/μL at the time of fever occurrence.
Initial data harvest from the electronic medical record (EMR) used International Classification of Disease (ICD) diagnoses codes. Encounters with ICD diagnosis of malignancy with ICD diagnosis of fever and/or neutropenia in patients aged 0-18 years were identified for evaluation. Included patients met the following criteria: age 0-18 years with oncologic diagnosis, single reported or documented temperature ≥ 38 degrees Celsius, ANC ≤ 500 cells/μL, presence of central line, and admission for IV antibiotics. Patients with acute myeloid leukemia, acute promyelocytic leukemia, or stem cell transplantation (SCT) were excluded. No lymphoma patients were excluded. Outlying causes for prolonged admission (scheduled chemotherapy with subsequent development of FN and ongoing admissions following adequate bone marrow recovery for other medical care or treatment of diagnosed infection) were excluded. Oncologic diagnoses included in the study after data harvest were B cell acute lymphoblastic leukemia (ALL), T cell ALL, Hodgkin lymphoma, Burkitt lymphoma, osteosarcoma, Ewing sarcoma, rhabdomyosarcoma, neuroblastoma, medulloblastoma, glioma, pituitary germinoma, retinoblastoma, clear cell carcinoma, small cell ovarian carcinoma, undifferentiated sarcoma, and malignancy of unknown origin. The ALL and lymphoma group for statistical analysis included the diagnoses of B cell ALL, T cell ALL, Hodgkin lymphoma, and/or Burkitt lymphoma.
Baseline data collection included age, oncologic diagnosis, sex, days since last chemotherapy, reported and measured temperature on admission, admission ANC, identified infections, antibiotics at discharge, discharge ANC, discharge APC, central line type, and granulocyte-colony stimulating factor (GCSF) use before, during, and/or after admission.
Continuous variables were compared with two sample t-tests. Wilcoxon ranked sum tests were used for continuous variables when the assumption of normality was not valid (ANC and APC). Categorical variables were analyzed with Chi-square tests and Fisher’s exact tests, as appropriate. Readmission and length of stay (LOS) between the groups were evaluated with Fisher’s exact test and t-test, respectively.
3. Results
3.1 -. Exclusion and Inclusion Data
Initial EMR data harvest yielded 510 potential patient encounters. Of 510 encounters, 98 (19.2%) met inclusion criteria. Group 1 (pre) consisted of 68/98 (69.4%) of included patient encounters. Group 2 (post) consisted of 30/98 (30.6%) of included patient encounters. Of the 412 excluded encounters, 153 (37.1%) were admissions that did not meet inclusion criteria or possessed cause for prolonged admission, 99 (24.0%) were incorrectly coded encounters, 77 (18.7%) were cases of fever not admitted, 59 (14.3%) were SCT patients, 4 (1%) were readmissions after an included patient encounter, and 20 (4.9%) were removed due to other exclusion criteria.
3.2 –. Baseline Group Data
Group 1 included 68 patient encounters, 41/68 (60.0%) with a diagnosis of ALL or lymphoma and 27/68 (40.0%) associated with other solid tumor malignancies. Group 2 included 30 patient encounters, 3/30 (10.0%) with a diagnosis of ALL or lymphoma and 27/30 (90.0%) associated with other solid tumor malignancies. Table 1 highlights baseline characteristics of Group 1 and Group 2. Significant differences between the groups were seen for ALL/lymphoma diagnoses, average age, average discharge APC, and use of GCSF. Increased use of GCSF in Group 2 correlates with the higher proportion of solid tumor malignancies.
TABLE 1.
Baseline characteristics
| Group 1 (n=68)a | Group 2 (n=30)a | P-valueb | |
|---|---|---|---|
| Oncological Diagnosis | |||
| ALL/lymphoma | 41 (60%) | 3 (10%) | <0.0001 |
| Other malignancy | 27 (40%) | 27 (90%) | |
| Average age in years | 10.4 (5.3) | 7.0 (5.0) | 0.004 |
| Gender | |||
| Male | 23 (34%) | 10 (33%) | >0.9 |
| Female | 45 (66%) | 20 (67%) | |
| Average days since last chemotherapy | 5.3 (4.4) | 7.0 (4.7) | 0.08 |
| Average reported temperature on admission (Celsius) | 38.5 (0.5) | 38.6 (0.7) | >0.9 |
| Average measured temperature on admission (Celsius) | 37.7 (0.8) | 37.7 (0.7) | >0.9 |
| Median ANC on admission (cells/μL) | 10 (0, 70) | 30 (0, 120) | 0.3 |
| Median ANC on discharge (cells/μL) | 495 (325, 840) | 465 (180, 680) | 0.1 |
| Median APC on discharge (cells/μL) | 916 (658, 1373) | 634 (468, 1065) | 0.04 |
| Central line type | |||
| Port | 49 (72%) | 25 (83%) | 0.4 |
| Broviac | 15 (22%) | 5 (17%) | |
| PICC | 4 (6%) | 0 (0%) | |
| GCSF use | |||
| Before admission | 15 (22%) | 20 (67%) | <0.0001 |
| During admission | 21 (31%) | 19 (63%) | <0.0001 |
| Continued after discharge | 11 (16%) | 19 (63%) | <0.0001 |
Reported as n (%), mean (SD), or as median (IQR).
From t-test (continuous variables), chi-square test (categorical variables), or Wilcoxon ranked sum test (continuous variables without normal distribution) comparing differences among groups.
ALL, acute lymphoblastic leukemia; ANC, absolute neutrophil count; APC, absolute phagocyte count; PICC, peripherally inserted central catheter; GCSF, granulocyte-colony stimulating factor; SD, standard deviation; IQR, interquartile range.
3.3 –. Outcomes Data
Within 14 days of discharge, 2 (3.0%) readmissions occurred in Group 1, and 1 (3.0%) readmission occurred in Group 2, P >0.9. There were no patient deaths in either group. Group 1 had average LOS of 7.43 days (SD=3.68 days) with a significantly shorter average LOS in Group 2 of 4.60 days (SD=2.42 days), P<0.0001. Group 1 had significantly more patients discharged with antibiotics (N=22, 32.0%) compared to Group 2 (N=4, 13.3%), P=0.05. There was no significant difference, however, in the number of confirmed infections between the groups as shown in Table 2.
TABLE 2.
Outcome data
| Group 1a | Group 2a | P-valueb | |
|---|---|---|---|
| Readmission within 14 days | 2 (3%) | 1 (3%) | >0.9 |
| Average LOS (days) | 7.43 (3.68) | 4.60 (2.42) | <0.0001 |
| Central line removal | 0 (0%) | 0 (0%) | - |
| Patient deaths | 0 (0%) | 0 (0%) | - |
| Confirmed infections | 33 (49%) | 9 (30%) | 0.09 |
| Antibiotics at time of discharge | 22 (32%) | 4 (13%) | 0.05 |
Reported as n (%), or mean (SD).
Readmissions from Fisher’s exact test, average LOS from t-test, and confirmed infections and antibiotics from chi-square test.
LOS, length of stay; SD, standard deviation.
4. Discussion
This data demonstrate that patients discharged with ANC ≥ 100 cells/μL and APC ≥ 300 cells/μL (Group 2) have no increased risk for readmission or mortality within 14 days of discharge compared to those discharged with ANC ≥ 500 cells/μL (Group 1). Villanueva et al published a series in 2016 of 299 FN discharges without count recovery criteria required for discharge. Readmission rate was 16.7% with 42% of the readmissions diagnosed with a significant infection.7 A subsequent publication by Campbell et al in 2018 analyzed 350 FN discharges with a retrospective analysis of count recovery subgroups by ANC. Discharges with ANC ≥ 100 cells/μL had similar safety outcomes to discharges with a higher ANC, with a trend toward increased readmissions with a discharge APC <500 cells/μL.4
This pilot study is limited as a retrospective review from a single center with a small evaluable sample size. Further limitations include inherent issues with EMR case identification by coding, and a smaller proportion of ALL/lymphoma patients in Group 2. However, unlike the Campbell et al publication, we utilized prospective count recovery criteria for discharge that should decrease additional bias based on individual physician practices. This data suggest inclusion of APC as part of count recovery criteria may add an additional measure of safety against readmission. This is a particularly important consideration for our patient population in a state with 63% rural counties, where transportation and access for readmission can be more difficult.9
Additionally, patients discharged with ANC ≥ 100 cells/μL and APC ≥ 300 cells/μL had a meaningful shorter LOS of 2.83 days compared to patients discharged with ANC ≥ 500 cells/μL. Pediatric oncology patients have been found to account for high volumes of healthcare utilization and spending.10,11 Shorter LOS can directly impact quality of life and healthcare costs. Results should be replicated in a larger multi-institutional study with attention to larger proportions of ALL/lymphoma patients as implementation of these discharge criteria may safely lower healthcare costs and improve quality of life for a highly vulnerable patient population.
Acknowledgements
Supported in part by the Biostatistics Shared Resource, Hollings Cancer Center, Medical University of South Carolina (P30 CA138313). The authors would also like to thank all faculty, trainees, and staff that implemented these criteria at MUSC.
Footnotes
Conflict of Interest
The authors declare that there is no conflict of interest.
References
- 1.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(4):e56–93. 10.1093/cid/cir073 [DOI] [PubMed] [Google Scholar]
- 2.Taplitz RA, Kennedy EB, Bow EJ, et al. Outpatient management of fever and neutropenia in adults treated for malignancy: American Society of Clinical Oncology and Infectious Diseases Society of America clinical practice guideline update. J Clin Onc. 2018;36(14):1443–1453. [DOI] [PubMed] [Google Scholar]
- 3.Lehrnbecher T, Robinson P, Fisher B, et al. The management of fever and neutropenia in children with cancer and hematopoietic stem-cell transplantation recipients: 2017 update. J Clin Onc. 2017;35(18):2082–2094. [DOI] [PubMed] [Google Scholar]
- 4.Campbell ME, Friedman DL, Dulek DE, et al. Safety of discharge for children with cancer and febrile neutropenia off antibiotics using absolute neutrophil count threshold values as a surrogate marker for adequate bone marrow recovery. Pediatr Blood Cancer. 2018;65:e26875. 10.1002/pbc.26875 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Bash RO, Katz JA, Cash JV, et al. Safety and cost effectiveness of early hospital discharge of lower risk children with cancer admitted for fever and neutropenia. Cancer. 1994;74(1):189–196. [DOI] [PubMed] [Google Scholar]
- 6.Pizzo PA. Management of fever in patients with cancer and treatment-induced neutropenia. N Engl J Med. 1993;328(18):1323–1332. [DOI] [PubMed] [Google Scholar]
- 7.Villanueva MA, August KJ. Early discharge of neutropenic pediatric oncology patients admitted with fever. Pediatr Blood Cancer. 2016;63(10):1829–1833. [DOI] [PubMed] [Google Scholar]
- 8.Griffin TC, Buchanan GR. Hematologic predictors of bone marrow recovery in neutropenic patients hospitalized for fever: implications for discontinuation of antibiotics and early discharge from the hospital. J Pediatr. 1992;121(1):28–33. [DOI] [PubMed] [Google Scholar]
- 9.Blueprint for Health. South Carolina Office of Rural health website. https://scorh.net/blueprint-for-health/.Accessed May 10, 2020.
- 10.Lekshminarayanan A, Bhatt P, Linga VG, et al. National trends in hospitalization for fever and neutropenia in children with cancer, 2007-2014. J Pediatr. 2018;202:231–237. [DOI] [PubMed] [Google Scholar]
- 11.Mueller EL, Hall M, Berry JG, et al. Healthcare utilization and spending by children with cancer on Medicaid. Pediatr Blood Cancer. 2017;64:e26569. 10.1002/pbc.26569 [DOI] [PubMed] [Google Scholar]