Chemotherapy-Induced Neutropenia and Febrile Neutropenia in the US: A Beast of Burden That Needs to Be Tamed?

Abstract Neutropenia and febrile neutropenia (FN) are common complications of myelosuppressive chemotherapy. This review provides an up-to-date assessment of the patient and cost burden of chemotherapy-induced neutropenia/FN in the US, and summarizes recommendations for FN prophylaxis, including the interim guidance that was recommended during the coronavirus disease 2019 (COVID-19) pandemic. This review indicates that neutropenia/FN place a significant burden on patients in terms of hospitalizations and mortality. Most patients with neutropenia/FN presenting to the emergency department will be hospitalized, with an average length of stay of 6, 8, and 10 days for elderly, pediatric, and adult patients, respectively. Reported in-hospital mortality rates for neutropenia/FN range from 0.4% to 3.0% for pediatric patients with cancer, 2.6% to 7.0% for adults with solid tumors, and 7.4% for adults with hematologic malignancies. Neutropenia/FN also place a significant cost burden on US healthcare systems, with average costs per neutropenia/FN hospitalization estimated to be up to $40 000 for adult patients and $65 000 for pediatric patients. Evidence-based guidelines recommend prophylactic granulocyte colony-stimulating factors (G-CSFs), which have been shown to reduce FN incidence while improving chemotherapy dose delivery. Availability of biosimilars may improve costs of care. Efforts to decrease hospitalizations by optimizing outpatient care could reduce the burden of neutropenia/FN; this was particularly pertinent during the COVID-19 pandemic since avoidance of hospitalization was needed to reduce exposure to the virus, and resulted in the adaptation of recommendations to prevent FN, which expanded the indications for G-CSF and/or lowered the threshold of use to >10% risk of FN.


Aim
Using the available literature, a review was undertaken to evaluate and summarize reports of the burden of neutropenia/ FN on patients with cancer (in terms of hospitalization and mortality) and US healthcare systems (in terms of cost) over the past 5 years. Following this, we reviewed recommendations from ASCO, ESMO, and NCCN for G-CSF prophylaxis of FN as well as the interim guidance from these societies on how FN management was adapted during the COVID-19 pandemic.

Search Strategy
A PubMed search was conducted. The search strategy comprised the following string: ("neutropenia" OR "febrile neutropenia" OR "neutropenic fever") AND ("United States" OR "America") AND ("infection" OR "antibiotic" OR "hospital*" OR "dose reduction" OR "dose delay" OR "dose intensity" OR "relative dose intensity" OR "RDI" OR "quality of life" OR "QoL" OR "cost" OR "economic" OR "afford*" OR "resource" OR "burden"). All terms were mapped to medical subject heading terms. Eligible reports included US-based real-world studies (observational, retrospective, and prospective studies) and healthcare resource utilization/cost analyses. The search was restricted to full English-language articles published within the last 5 years (since January 1, 2015). In addition, a supplementary PubMed search was conducted to identify reports on the mortality of FN in patients with hematologic malignancies; only one eligible publication was identified.
Guidelines reviewed comprised the ASCO 2015 recommendations for the use of white blood cell growth factors, ESMO 2016 clinical practice guidelines on the management of FN, and NCCN guidelines on hematopoietic growth factors (version 4, 2021). 1,8,16 Recommendations from these 3 societies in response to the COVID-19 pandemic were also reviewed. [27][28][29]

Included Studies
Overall, 88 papers were identified. After title and abstract screen, 74 papers were excluded due to not being conducted in the US (n = 31), not reporting burden (hospitalization, mortality, or cost; n = 17), no mention of neutropenia/FN (n = 10), not in patients with cancer (n = 10), neutropenia/FN not caused by myelosuppressive chemotherapy (n = 5), and being a non-human study (n = 1). This left 14 eligible papers.

Hospitalizations
Eleven studies published in the last 5 years reported findings on hospitalization rates and length of stay (LOS; Table 1). Overall, these papers analyzed over 670 000 and 118 000 hospitalizations or discharges for neutropenia/FN in children and adults, respectively.
For patients with cancer who visit the emergency department (ED) for FN, nearly all visits end in hospitalization. Among pediatric patients, it has been reported that 82.3% of visits resulted in hospitalization, 36 increasing to 94.0% in one study of patients of any age. 30 Among pediatric patients with cancer, 10.1%-22.7% of hospitalizations were due to neutropenia/FN. 32,34,35 Patientrelated factors found to increase the likelihood of hospitalization included younger age (especially <10 years old), male, non-Hispanic white ethnicity, and having private rather than public health insurance. 31,32,34,35 Disease-related risk factors consisted of primary diagnosis of acute lymphoblastic leukemia and higher numbers of chronic conditions. 32,34,35 Moreover, a higher risk of hospitalization was observed in patients admitted through the ED, and those presenting to institutions in the Midwest or Western regions. 31,32 The finding that younger children are at a higher risk of hospitalization reflects that hematologic malignancies, especially leukemias, are typically diagnosed in young children and in these patients bone marrow suppression is caused not only by chemotherapy but also by the cancer itself originating in the bone marrow. 32,35 Indeed, hospitalization for FN was found to be common among pediatric patients with acute lymphoblastic leukemia. 34,35 The observation that hospital admission for neutropenia/FN was more common through the ED also aligns with the urgent need to immediately treat patients with broad-spectrum antibiotics to try and combat the causative infectious agent. 32 Moreover, the finding that hospitalization for FN is more likely to occur in institutions based in Midwestern or Western regions could be due to a number of factors including differences in the management protocols within institutions and/or healthcare systems, as well as factors such as patient travel distance to hospital and ease of access. 34,35 Table 1.

Continued
Among adults, one retrospective analysis of over 90 000 hospitalizations estimated that neutropenia accounted for 5.2% of all cancer-related hospitalizations. 32 In studies of adult patients with (AML)/myelodysplastic syndromes (MDS) or diffuse large B-cell lymphoma (DLBCL), 65.7% and 34.5% of hospital admissions were due to FN, respectively. 38,39 A number of patient-and disease-related factors reported to increase the likelihood of hospitalization for neutropenia/FN were in-line with those listed in FN management guidelines 1,8,16,40 such as older age, presence of comorbidities, advanced stage disease, and first cycle chemotherapy average relative dose intensity ≥85%. 30,32,39 Although, one study reported that younger age (18-44 years) was associated with an increased risk of hospitalization. 32 This finding may have been influenced by other characteristics of the patient population (eg, cancer type and number of chronic conditions). Baseline sarcopenia was an additional factor found to increase the risk of hospitalization for FN. 39 However, the older age of sarcopenic versus non-sarcopenic patients (mean, 68.1 vs 61.2 years) may have contributed to this finding. 39 Findings on type of insurance were conflicting with one study reporting that private insurance was associated with hospitalization, 32 whereas another found that hospitalization was more likely among those with public insurance. 30 Increased likelihood of hospitalization with public insurance is expected due to less resources available for outpatient management. Similar to pediatric patients, admittance from the ED was again associated with increased risk of hospitalization for FN in adult patients. 32 One study reported a higher likelihood of hospitalization in urban teaching hospitals 32 whereas another found that hospitalization was more common in metropolitan non-teaching or non-metropolitan hospitals. 30 Again, these differences are potentially due to variations in the management protocols used and patient-related factors.
In elderly patients (aged ≥66 years), inpatient hospital care was needed for 2121 of 2407 (88.1%) cases of FN in patients with breast cancer, 3571 of 3840 (93.0%) in patients with lung cancer, and 3342 of 3587 (93.2%) in patients with non-Hodgkin lymphoma (NHL). 37 A larger proportion of patients with NHL (28.3%) and lung cancer (16.7%) was aged ≥80 years than those with breast cancer (9.9%), and patients with breast cancer had fewer comorbid conditions than those with lung cancer or NHL, possibly contributing to the higher rates of inpatient hospital care. 37 The average LOS of patients admitted to hospital for FN was 4.0-8.5, 5.7-9.6, and 6.2-6.8 days for children, adults, and elderly patients, respectively. 20,[31][32][33][34][35]37 The shorter LOS among pediatric and elderly patients is likely due to these patients often being admitted on a precautionary basis and once observed briefly are then dismissed. Although for elderly patients this could also be due to higher mortality rates.
Longer LOS in children was associated with both younger (<1-4 years) and older (10-19 years) age, non-Caucasian race, and having public insurance rather than private insurance. 31,33,34 The finding regarding public insurance was speculated to be due to pressure on hospitals to discharge patients quickly when the patients' insurance is private due to having to justify the stay. 33 Pediatric patients with leukemia, bone tumors, or soft tissue sarcomas, as well as comorbid infections or hypotension had longer LOS. [31][32][33] Longer LOS was also common among adult patients with leukemia. 32 Regarding institution-related factors, one Californian retrospective analysis found that median LOS was significantly longer at pediatric cancer specialty centers compared with non-specialty centers (9 vs 7 days; P < .0001). 33 The authors hypothesized that this may reflect the increased severity of illness seen at specialist centers as well as the increased likelihood that patients at specialist centers may initiate their next cycle of chemotherapy while still an inpatient once their neutrophil counts have returned to normal. 33

Mortality
Eight studies, 5 in children with cancer, 2 in adults with solid tumors, and 1 in adults with hematologic malignancies, reported in-hospital mortality rates (Table 2). Among pediatric patients with cancer hospitalized for neutropenia/FN, mortality rates ranged from 0.4% to 0.8%, 31,32,34,35 with one study reporting a higher rate of 3%. 33 A significantly higher risk of mortality was observed among adolescents aged 15-19 years versus children aged 0-4 years, and in patients with infections (eg, mycosis, meningitis, pneumonia, and sepsis). 31 The authors hypothesized that the disease biology may be different in adolescent compared with younger patients, contributing to unfavorable outcomes. 31 Other factors which potentially could have contributed to greater mortality included variation in care, social support, delays in seeking care, and socioeconomic status. 31 Among adults with cancer hospitalized for neutropenia/FN, mortality rate ranged between 2.6% and 7.0% for those with solid tumors 20,42 and was 7.4% in a retrospective chart review of patients with hematologic malignancies. 43 Risk factors for mortality among patients with solid tumors included older age (≥65 vs <65 years), lung cancer versus other solid tumors, the presence of comorbidities, infection, sepsis, or pneumonia, and admission to intensive care. 20,42 The higher mortality for patients with lung cancer is likely due to these patients being older than patients with other solid tumors, with 50.0% and 31.6% of patients aged ≥65 years, respectively. 42 In addition, patients with lung cancer were more likely to have 2 or more comorbidities, and comorbid heart and lung diseases, than patients with other solid tumors. 42 The presence of sepsis or elevated bilirubin were risk factors for mortality among patients with hematologic malignancies. 43 An additional retrospective case-control study in elderly patients (aged ≥60 years) with AML reported a significantly longer duration of grade III/IV neutropenia with clofarabine-based induction, which was hypothesized to contribute to a 7-times higher 30-day induction mortality rate compared with fludarabine, cytarabine, and G-CSF (FLAG). 41 Table 3 summarizes the findings on the cost of neutropenia/FN care in patients with cancer; 4 of the 6 studies were for treatment of pediatric patients. In an analysis of 2012 data from the National Inpatient Sample, total costs of hospitalizations for neutropenia among adult patients with cancer amounted to $2.3 billion. 32 In pediatric patients, 2 cross-sectional analyses of the KID demonstrated that the financial burden of FN care rose from $587 million in 2009 to $881 million in 2012. 34,35 However, another study using 2012 KID data estimated total costs to only be $439 million. 32 This difference in costs despite using the same database is likely due to methodological differences such as the age of included patients (≤17 vs ≤19 years), how cancer-related neutropenia hospitalizations were identified (ICD-9-CM based on previous literature vs clinical classification software codes), and how costs were reported (2012 USD vs conversion from 2012 USD to 2015 USD dollar using the consumer price index). 32,34 The mean cost per FN hospitalization was $20 000-40 000 for adults 20,32 and $8000-65 000 for children 31,32,34,35 with cancer.

Burden on Healthcare Systems
Factors significantly associated with higher costs for pediatric patients included older age (10-19 years), non-private insurance payers, and comorbid infection. 31 In one study, a primary diagnosis of leukemia in both adults and children was associated with a higher mean cost per neutropenia hospitalization compared with other cancer types. 32 Within this study, patients with leukemia had the longest mean LOS compared with other cancer types, potentially contributing to the higher costs of neutropenia care. 32 Among elderly patients, mean total costs of care per FN episode were reported to be highest in patients with NHL (~$15 000) and lowest for those with breast cancer (~$12 000). 37 The higher cost of  FN care for patients with NHL was again potentially due to these patients having the longest mean LOS compared with patients with breast or lung cancer. 37 Accounting for all but 1% of mean total costs, inpatient care was the largest driver of cost for FN care. 37 Costs associated with outpatient-only care for FN were far lower than costs for inpatient care ($849 for breast cancer, $841 for lung cancer, and $1322 for NHL) but care in this setting occurred much less frequently. 37

FN Prophylaxis with G-CSF in Patients with Cancer: Current Guidelines
Evidence-based guidelines from societies based in the US (ASCO and NCCN) and Europe (ESMO) align in their recommendations and advise that G-CSFs, such as filgrastim and pegfilgrastim, should be used as prophylaxis for FN in patients with cancer who are at high risk of FN and receiving myelosuppressive chemotherapy. 1,8,16 G-CSF is recommended as primary prophylaxis when a chemotherapy regimen carries a high risk (≥20%) of FN and in patients with an intermediate risk (10-20%) if they have ≥1 risk factor. 1,8,16 Presently, the guidelines generally do not cover targeted agents. These newer agents have more precise mechanisms of action than chemotherapy but many are still myelosuppressive, including monoclonal antibodies (eg, rituximab), immunomodulatory drugs (eg, lenalidomide), and kinase inhibitors (eg, cyclin-dependent kinase 4/6 inhibitors), many of which are used in metastatic or relapsed/refractory disease. 44 In the May 2021 release of the NCCN guidelines on hematopoietic growth factors, for a number of regimens recommendations for G-CSF support apply with or without the addition of monoclonal antibodies (eg, trastuzumab, rituximab). 8 There is the potential for increased neutropenia risk with the addition of monoclonal antibodies; for example, rituximab has been associated with prolonged neutropenia with or without chemotherapy. 8,45

FN Prophylaxis with G-CSF in Patients with Cancer: COVID-19 Considerations
In response to the COVID-19 pandemic, ASCO, ESMO, and NCCN released guidance on G-CSF administration to reduce the risk of drug-induced neutropenia/FN and associated infection in patients with cancer, and ensure that their treatment can be delivered at the most effective dose and on time (Table 4). [27][28][29] All 3 interim guidelines recommended that the indications for G-CSF were expanded and/or the threshold of G-CSF use was lowered to >10% risk, to reduce the risk of FN and the numbers of patients requiring treatment in hospitals or emergency settings, thereby decreasing COVID-19 exposure. [27][28][29] These statements were concordant with pre-COVID-19 recommendations advocating the use of G-CSF prophylaxis in special circumstances for patients at intermediate risk of FN (10%-20%). 1,8 The NCCN recommended that self-administration of filgrastim or long-acting pegfilgrastim, or on-body injector pegfilgrastim should be considered to reduce the risk of contracting COVID-19 during outpatient visits. 29 In patients with potential FN, ASCO recommended that the neutropenic status of patients should be first evaluated by telemedicine or phone to help decide if the patient should be assessed in the clinic or sent to the ED. 27 In patients with confirmed FN, rapid COVID-19 testing should be used, if available, to inform the appropriate location for continued care and the level of personal protective equipment needed for caregivers. 27 ESMO noted that the risk of acute respiratory failure due to G-CSF-induced leukocyte recovery in patients with COVID-19-associated lung infections did not outweigh the benefits of using G-CSF. 28 Expanded use of antibiotics was recommended, alongside G-CSF, because of the risk of a delay to emergency visits during the pandemic for patients with fever. 28 Regimens that are unlikely to induce FN were advised for patients with solid tumors undergoing non-curative treatment. 28

Discussion
This review of recent real-world evidence from the US indicates that chemotherapy-induced neutropenia/FN place a significant burden on patients with cancer, in terms of hospitalizations and mortality, and on healthcare systems, in terms of cost. Most patients presenting to the ED with neutropenia/ FN will be hospitalized, 30,36,37 with the average LOS being 6 days for elderly patients, ~8 days for children, and up to 10 days for adults. 20,[32][33][34]37 In this review, mortality rates for children with cancer hospitalized for neutropenia/FN were 0.4%-3.0%, [32][33][34][35] with rates of 2.6%-7.0% reported in adults with solid tumors 20,42 and 7.4% in adults with hematological malignancies. 43 Improved awareness of factors that increase mortality during FN-related hospitalization (eg, older age and infection 20,31,42 ) may help healthcare professionals improve survival by allowing earlier identification of those at risk and provide more individualized care. 42 As well as directly affecting patient outcomes, neutropenia/ FN place a significant cost burden on healthcare systems in the US and worldwide. In 2012, total hospitalization costs for neutropenia/FN in the US amounted to >$2 billion for adults and up to $880 million for children with cancer. 32,34 Mean cost per hospitalization was reported to be up to $15 000 for elderly patients (2007-2015 data), 37 $40 000 for adults (2009-2011 data), 20 and $65 000 for children (2012 data). 34 While restricting our search to articles published within the past 5 years to give contemporary estimates of FN, a limitation of this review is that several papers focus on population databases with data gathered between 2005 and 2015. Differences in methods between studies, such as patient classification ages for pediatric and elderly patients, makes comparing data and drawing conclusions more complex. Furthermore, due to a lack of a single discharge code for FN, several of the included studies identified cases of neutropenia/FN in patients with cancer using combinations of diagnostic codes, which may be subject to error, identify patients with neutropenia caused by means other than chemotherapy, and select a sample biased toward sicker patients who are more likely to be hospitalized. Some of the databases used in the included studies do not track revisits and therefore some patients may have been counted more than once. Although providing insight into direct costs of care for neutropenia/FN, the included studies did not assess societal costs including productivity loss and patient and caregiver time and transportation costs. However, strengths of the included studies include their representation of real-world clinical practice and inclusion of large numbers of patients, providing an updated overview of the burden of neutropenia/ FN on healthcare systems and patients with cancer.
Efforts are needed to decrease hospital stays, thereby reducing the burden of neutropenia/FN on patients and healthcare systems. Outpatient FN treatment was found to be far less costly than inpatient care, but also far less common. 37 Outpatient management of FN can be optimized through the knowledge and experience of physicians and pharmacists, who can identify suitable patients for oral antibiotic therapy and monitor treatment. 46,49 The development of educational initiatives for patients with cancer, their caregivers, and healthcare professionals can also decrease the burden of neutropenia/FN by informing them on how to reduce the risk of neutropenia-related infections and enabling them to proactively treat infection. 50 This can be provided by physicians, pharmacists, oncology nurses, and physician assistants, who can also contribute to the development of and adherence to their institution's guidelines and processes. 49,51 FN management can also be improved through the design and implementation of more efficient processes, to overcome barriers preventing the timely administration of antibiotics. 1,52,53 The Multinational Association for Supportive Care in Cancer (MASCC) risk index score was developed to identify patients with FN at low risk of serious complications or death, and guidelines recommend that it can be used to identify patients that can be managed in the outpatient setting. 1,7,47 However, application of MASCC score does not appear to be resulting in fewer hospitalizations. Real-world experience indicates that the MASCC score is too slow to be used in real-life emergency medicine and is inaccurate, with a predictive value of only 83%. 30,54 The Clinical Index of Stable Febrile Neutropenia (CISNE) is a newer score found to be more accurate than MASCC and more appropriate for use in the emergency setting. [55][56][57] However, CISNE only applies to solid tumors and is not recommended for use in patients presenting with septic shock or a previously known severe infection. Continued research is required to develop and validate risk stratification tools that can quickly and accurately identify patients at highest risk of complications from FN in real time in clinical practice, while being mindful of cost-effectiveness and patient quality of life. G-CSF is not recommended due to the potential for increasing pulmonary inflammation and inflammatory cytokine (eg, IL-6) production associated with severe COVID-19 infection • Self-administration of daily filgrastim or long-acting pegfilgrastim (1-3 days after chemotherapy) or use of an on-body injector pegfilgrastim are recommended to minimize visits to outpatient centers and reduce the risk of COVID-19 exposure ESMO 28 • For patients with solid tumors not treated for cure, consider administering regimens at low risk of FN • For use of regimens with a higher risk of FN, there must be considerable evidence that clearly outweighs potential emergency intervention and COVID-19 exposure • G-CSF indication should be expanded to include patients receiving chemotherapy with a lower risk of FN (the theoretical concern of acute respiratory failure due to G-CSF-induced leukocyte recovery in patients with COVID-19 pulmonary infection does not outweigh the benefit); however, this approach may require additional outpatient clinic visits • For outpatient management of FN in patients with lower risk, well-documented and verified criteria are available (eg, the MASCC FN risk group stratifications 47 ), with published randomized trials using oral antibiotics • Use of antibiotic prophylaxis and/or prescription of stand-by antibiotics should be expanded due to a potential risk of a delay to emergency visits for patients who develop fever • Critical review and reduction of the use of steroids is recommended, if possible • In patients receiving a fluoropyrimidine, genetic testing to identify patients with DPD deficiency is recommended 48 • No evidence is currently available demonstrating that neutropenia due to PARP or CDK4/6 inhibitors results in an increase in associated viral infections Evidence-based guidelines recommend that G-CSFs are used as prophylaxis for FN in patients with cancer receiving a high-risk (≥20%) chemotherapy regimen, or an intermediate-risk (10%-20%) regimen if ≥1 risk factor is present. 1,8,16 In response to the COVID-19 pandemic, ASCO, ESMO, and NCCN released guidance on G-CSF administration to reduce the infection risk in patients with cancer at risk of FN. The recommendations suggested that the indications for G-CSF were expanded and/or the threshold of G-CSF use was lowered to >10% risk of FN. [27][28][29] Guidelines recommend that filgrastim is administered subcutaneously at a dose of 5 μg/kg/day starting 1-3 days after chemotherapy until recovery of neutrophils to normal levels. 1,8,16 Pegfilgrastim should be given subcutaneously as a single 6 mg dose the day after myelosuppressive chemotherapy. 1,8,16 When taken according to guideline recommendations, filgrastim and pegfilgrastim are equally effective. 58 However, in clinical practice filgrastim is often under-dosed and therefore pegfilgrastim demonstrated more effectiveness than filgrastim in terms of reducing the incidence of FN and hospitalization, and achieving target dose intensity of chemotherapy. 58 The most common adverse event with G-CSF therapy is bone pain, and patients should be encouraged to report this, together with any other adverse effects, to their treatment team. 1,8,16 Usually, bone pain is managed with standard analgesics including acetaminophen and nonsteroidal anti-inflammatory drugs. 1,8,16 Despite the use of prophylactic G-CSF and other management strategies, the impact of chemotherapy-induced myelosuppression on patients remains significant. 59 Improving communication between healthcare professionals and their patients is crucial to bettering patients' understanding of chemotherapy-induced myelosuppression and encouraging shared decision-making in regards to treatment. 59 Biosimilars provide lower-cost alternatives to their reference medicines. 60,61 Use of G-CSF biosimilars may improve costs of care and help reduce the burden of neutropenia/FN to healthcare systems. 62

Conclusion
In the US and indeed globally, chemotherapy-induced neutropenia/FN remain a significant burden on patients with cancer in terms of hospitalization and mortality, and on healthcare systems in terms of cost. This burden may be reduced through efforts to decrease the number and duration of hospital stays via the optimization of outpatient care, efficiency improvements in management processes, development of educational initiatives, and design of risk-stratification tools validated in clinical practice.