Prognostic indicators predictive of chemotherapy-induced febrile neutropenia outcomes in adult cancer patients: a systematic review.

Abstract

Review question/objective What are the prognostic indicators predictive of chemotherapy-induced febrile neutropenia outcomes among adult cancer patients? Background Neutropenia is one of the commonest adverse effects of myelosuppressive chemotherapy. Defined as a reduction of circulating neutrophils in the blood caused by administration of chemotherapeutic agents,1 it affects more than 70% of cancer patients undergoing systemic chemotherapy2 but varies with the number of doses administered and chemotherapy regimen.3, 4 With the increased number of cancer diagnoses and patients undergoing treatment especially among the aging population, a rise in the incidence of Chemotherapy-Induced-Neutropenia (CIN) is expected.5, 6 Compared with other adverse effects, CIN and its associated complications are a significant cause of morbidity and mortality, and have a negative impact on overall survival of cancer patients7 and healthcare expenditure for cancer care.8-10 CINs pose a substantial risk of life-threatening infections to cancer patients which range between 10%-50% for solid tumour patients and more than 80% for patients with haematology malignancy with a general mortality rate between 7-20%.6, 11 However, with the development of newer cytotoxic agents and dose-dense chemotherapy schedules for a variety of tumour types that suggest improved survival rates, patients are faced with more severe and prolonged neutropenia.12 Both degree (especially when absolute neutrophil concentration is less than 0.5x109/L) and duration (more than one week) of neutropenia are well recognised factors for more serious infections that can lead to higher rates of morbidity and mortality.13, 14 Febrile Neutropenia and its management Fever is the classical presentation of infection in patients with neutropenia. At the onset of Febrile Neutropenia (FN), the standard of care involves immediate medical assessment, blood tests and cultures, administration of intravenous (IV) broad-spectrum antibiotics and hospitalization for further observation.15, 16 It is not uncommon for FN patients' medical condition to deteriorate rapidly within hours from the onset of fever and a mild infection may easily progress to sepsis syndrome requiring admission to the intensive care unit (ICU).17, 18 Because of the reported morbidity and mortality associated with infection in FN patients, FN episodes are universally managed as medical emergencies and patients are required stay in the hospital for IV antibiotics until fever and any active infection are resolved; and neutrophils count recovery.19, 20 The mean length of stay for these patients is 11.5 days and the mean cost of hospitalization for each episode of infection is estimated to be USD$19,000.8, 21 Consequences and Preventive Measures for Febrile Neutropenia and its complications Other consequences of FN include delays in treatment and dose reductions for subsequent cycles of chemotherapy, both of which can compromise disease response and overall survival.22, 23 Intravenous antibiotics therapy results in prolonged hospital stay for patients and exposes them to increased risk of multi-drug resistant nosocomial infection.24 During hospitalisation, they are nursed in a protective environment that results in social isolation and interruption in activities of daily living, in addition to having a diet restrictions imposed on them.25 In short, the consequences of FN do not only affect clinical outcomes for patients but the preventive measures also have an impact on the their quality of life.26, 27 Given these circumstances, numerous strategies have been implemented to prevent and minimise FN, and its complications. To shorten the duration or decrease the severity of neutropenia, prophylactic use of colony stimulating factors (CSFs) for FN patients with high risk for complications has been recommended.28-32 It has also been shown that the use of CSFs allows clinicians to achieve dose intensity for some chemotherapy regimens and keep the patients on schedule for each cycle of treatment.33, 34 Other approaches include the efficacious use of prophylaxis antibiotics35-38 and prompt initiation of treatment at the onset of FN.39 In the recent years, the focus of research studies on FN management have been aimed at enhancing the delivery of safe and effective care by exploring risk-based antibiotic therapy40-44; optimizing the utilisation of healthcare resources and improving the quality of life for patients and care-givers.45, 46 Some studies looked at the epidemiology and microbiological trends of infections associated with FN so that antibiotic therapy can be optimized;47-49 while others are related to developing more specific diagnostic tools for early detection of infection and clinical deterioration. 50-54 However there has been a huge amount of interest in risk assessment and, prognostic models in predicting neutropenia occurrence 55-57 as well as FN outcomes.40, 42, 58 This could possibly be due to the observation that patients with FN are a heterogeneous group, and not all FN patients are required to receive the standard management.59, 60 In order to address the change, a valid and reliable assessment tool is needed to categorise patients into specific risk groups so that treatment can be tailored accordingly.40, 42, 61 Prognostic indicators in Chemotherapy-induced Febrile Neutropenia outcomes The utilisation of assessment tools to measure risk and predict treatment related outcomes is a common practice in the management of patients with cancer. In the form of prognostic models, these tools are used to assist health practitioners in clinical decision making by providing information on the likelihood of disease recurrence or occurrence of a treatment related outcome.62 In the aspect of FN management, some of the widely used prognostic models that predict the outcome of FN episodes by classifying patients into low and high risk groups are the Talcott classification model40 and Multinational Association Supportive Care for Cancer (MASCC) scoring system.42 These models were developed based on a combination of significant patients' characteristics and variables in influencing the outcomes of FN, which were derived from clinical expertise and observational studies. Even though these models have been validated in numerous studies and reported to be reliable, issues such misclassification for the low risk group with low sensitivity to predict deterioration 58, 63 and limited discriminatory ability when used among the haematology group of cancer patients have been reported.64 These limitations could be related to the variables used to construct the prognostic models. Variables such as performance status, Absolute Neutrophil Count (ANC) at the onset of FN, duration of neutropenia and status of the bone marrow function have been reported to be significant indicators in the prognostication of FN outcomes, especially for haematological cancer patients13, 14, 58, 65 in the aspect of ANC recovery. However, these variable were not included as part of the scoring system hence affecting the generalisability of the scores in categorising the respective risk groups. Since the development of both prognostic models, a substantive amount of studies have been published on this aspect of care and the numbers of empirical studies on this aspect of care continue to grow. More prognostic indicators have been identified for both low and high risk groups66-68 even though reports of individual clinical variables and their significance in FN outcomes have been inconsistent due to small sample studies.69, 70 In addition, other prognostic models have been developed 43, 63, 71 and more recently the use of bio-markers to assist in early detection of bacteraemia has been reported.72-74 Given the lack of consensus on the indicators that accurately categorise and predict the outcome of FN episodes, there is a need for a systematic review to be conducted to examine the current available evidence for prognostic indicators predictive of FN outcomes in the assessment of cancer patients. The proposed systematic review will focus on adult patients with cancer who experience FN after receiving myelosuppressive systemic chemotherapy. The findings from this review will be analysed and presented in two different sections with a separate analysis of variables prognostic of high risk and low risk characteristics. Subgroup analysis will be performed for patients with solid tumour and haematology malignancy presented with FN because they are a heterogeneous group in relation to the recovery of ANC levels. Patients with haematology malignancies such as acute myeloid leukaemia often present with abnormal leukocytes and neutrophils in terms of the quantity and quality and these affect the recovery of ANC level.75 However this condition is not often present in solid tumour patients unless the tumour has spread to the bone marrow. This review will use the following definitions: (i) Fever - a single oral temperature measurement of ≥38.3°C (101°F) or a temperature of ≥38.0°C (100.4°F) sustained over a 1-hour period.76 (ii) Neutropenia - an ANC of <1000 cells/mm3 or an ANC that is expected to decrease to <500 cells/mm3 during the next 48 h.76 iii) Prognostic indicator - individual variables that strongly influence and predictive of the future occurrence of the outcomes of disease or condition regardless of therapy.77 In the context of this review, prognostic indicators will be used to represent clinical and laboratory variables which are associated with the favourable and unfavourable outcomes of FN episodes in adult cancer patients post chemotherapy. The primary outcome of this systematic review is the identification of prognostic indicators predictive of the clinical outcomes of FN episodes. Even though randomised controlled trials (RCT) are accepted as the gold standard in the conduct of interventional research studies, the majority of study designs for the identification of prognostic factors are likely to be cohort, observational, case-control, case-cohort and descriptive. This systematic review will report the best available evidence, based on both experimental and observational studies. Prior to undertaking the systematic review on this topic, a preliminary search from the Cochrane Library, Joanna Briggs Institute Library of Systematic Review, MEDLINE and the CRD Database of Abstracts and Review has been carried out to establish that no previous systematic reviews had been published on this topic. Inclusion criteria Types of participants The participants of interest include adults, 18 years old and above with cancer, who developed febrile neutropenia as a result of myelosuppressive systemic chemotherapy either in an in-patient or outpatient setting. Types of intervention(s)/phenomena of interest The focus of this systematic review is on the prognostic indicators predictive of chemotherapy-induced febrile neutropenia outcomes in adult cancer patients. Types of studies As the proposed systematic review is looking at evidence about prognostic factors, it will consider any experimental study design including randomised control trials; non-randomised controlled trials, quasi-experimental studies, cohort studies, case control studies, case-cohort studies and case series that examined prognostic indicators for chemotherapy-induced febrile neutropenia for adult patients post myelosuppressive systemic chemotherapy. Both retrospective and prospective studies (published or unpublished) that report sufficient data for analysis will be included. Types of Outcome(s) The primary outcome of this review is to identify the indicators predictive of FN outcomes (favourable or unfavourable) that influence the incidence of morbidity and mortality of adult cancer patients related to FN. Search strategy The search strategy aims to find both published and unpublished studies. A three-step search strategy will be utilised in this review. An initial limited search of MEDLINE and CINAHL will be undertaken followed by analysis of the free text and keywords contained in the title and abstract, and of the index terms used to describe article. A second search using all identified keywords and index terms will then be undertaken across all included databases. To accommodate for the differences in search features between databases, the search strategy will be modified accordingly. The reference lists of relevant reports and articles will be searched for additional studies. All studies published in English will be considered for inclusion in this review. Published reviews identified through the screening search that have addressed risk factors related to FN outcomes will also be retrieved in order to examine the included primary studies with relevance to the objectives and criteria for this systematic review. If full text articles or abstracts are not available the paper will be ordered through inter-library loans and screened for relevance. Databases to search will include the following databases from their inception up to December 2011: MEDLINE EMBASE CINAHL Cochrane Central Register of Controlled Trials (CENTRAL) Web of Science Science Direct Scopus Mednar Assessment of methodological quality Assessment of internal validity will be established based on the specific study design. JBI MAStARI contains standardised appraisal instruments for cohort, case control and case series studies. These are the designs most likely to be associated with prognostic research methods. This review will also consider randomised or pseudo randomised controlled trials in the event that any are identified. In each case, the relevant appraisal instrument from JBI MAStARI will be utilised to assess methodological quality. Critical appraisal will be undertaken by two independent reviewers following discussion of the criteria, their core definitions and how these should be applied. In the event that there are any disagreements, these will be resolved through discussion, or with a third reviewer. The specific appraisal instruments and checklist criteria are listed in Appendix I. Data collection Data from the studies included in the review will be extracted and assessed using the extraction forms from Joanna Briggs Institute Meta Analysis of Statistics Assessment and Review Instrument (JBI-MAStARI) for experimental and non-experimental studies respectively. (Appendix II) The data extracted will include specific details about the demographics, populations, study methods and outcomes of significance to the review question and specific objective. Data synthesis All results will be double entered in order to avoid data entry errors. Study results will be investigated using a fixed effect model as the preferred model, however, in the presence of statistical heterogeneity; a random effects model will be used to explore the heterogeneity. Statistical heterogeneity will also be explored via the I2 value and, sub group analysis where possible. Where statistical pooling is not possible the findings will be presented in narrative form. Conflicts of interest No known conflict of interest to report. Acknowledgements As this systematic review forms part of a submission of the award of a Masters of Clinical Science, a secondary reviewer will only be used for the critical appraisal stage.