YJINF3525_proof ■ 30 April 2015 ■ 1/9 Journal of Infection (2015) xx, 1e9
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
www.elsevierhealth.com/journals/jinf
Q8
Q7
Febrile neutropenia in children treated for malignancy Chris D. Barton a,c, Lucy K. Waugh a, Maryke J. Nielsen a, ´phane Paulus b,d,* Ste a
Department of Haematology & Oncology, Alder Hey Children’s Hospital, Eaton Road, Liverpool L12 2AP, United Kingdom b Department of Infectious Diseases, Alder Hey Children’s Hospital, Eaton Road, Liverpool L12 2AP, United Kingdom c Institute of Translational Medicine, University of Liverpool, Liverpool L69 3BX, United Kingdom d Institute of Global Health, University of Liverpool, Liverpool L69 3BX, United Kingdom Accepted 21 April 2015 Available online - - -
KEYWORDS Malignancy; Paediatrics; Febrile neutropaenia; Neutropenic sepsis
Summary Febrile Neutropenia (FN) in children treated for malignancy is a common and direct sequela of chemotherapy. Episodes of FN can be life-threatening, and demand prompt recognition, assessment and treatment with broad spectrum antibiotics. While in the majority of episodes no causal infection is identified, 10e20% are secondary to a bloodstream infection (BSI). A reduction in episodes of BSI could be achieved through robust infection prevention strategies, such as CVL care bundles. Alongside good antimicrobial stewardship, these strategies could reduce the risk of emergent, multi-drug resistant (MDR) infections. Emerging bacterial pathogens in BSI include Viridans Group Streptococci (VGS) and Enterobacteriaceae such as Klebsiella spp. which are known for their ability to carry MDR genes. There is also increased recognition of the role of invasive fungal infection (IFI) in FN, in particular with Aspergillus spp. Novel diagnostics, including multiplex blood and respiratory polymerase chain reaction assays can identify infections early in FN, facilitating targeted therapy, and reducing unnecessary antimicrobial exposure. Given appropriate, and sensitive rapid diagnostics, potential also exists to safely inform the risk assessment of patients with FN, identifying those at low risk of complication, who could be treated in the out-patient setting. Several clinical decision rules (CDR) have now been developed and validated in defined populations, for the risk assessment of children being treated for cancer. Future research is needed to develop a universal CDR to improve the management of children with FN. ª 2015 Published by Elsevier Ltd on behalf of The British Infection Association.
* Corresponding author. Alder Hey Children’s Hospital, Eaton Road, Liverpool L12 2AP, United Kingdom. Tel.: þ44 151 228 4811. E-mail addresses: [email protected] (C.D. Barton), [email protected] (L.K. Waugh), [email protected] (M.J. Nielsen), [email protected] (S. Paulus). http://dx.doi.org/10.1016/j.jinf.2015.04.026 0163-4453/ª 2015 Published by Elsevier Ltd on behalf of The British Infection Association. Please cite this article in press as: Barton CD, et al., Febrile neutropenia in children treated for malignancy, J Infect (2015), http:// dx.doi.org/10.1016/j.jinf.2015.04.026
61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122
YJINF3525_proof ■ 30 April 2015 ■ 2/9
2 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62
Introduction Children with cancer receiving cytotoxic chemotherapy and other antineoplastic therapy experience significant neutropaenia as a direct consequence of their treatment. The presence of an indwelling central line, with often concurrent loss of the mucosal integrity of the gastrointestinal tract, further predisposes these patients to bacteraemia and sepsis. Febrile episodes are observed in 34% of neutropenic periods in children treated for cancer, with bacteraemia identified in 10e20% of cases whereas an unknown aetiology is the most likely outcome (up to 79% of the cases).1
Definition and immediate management in children with febrile neutropenia The National Institute for Health and Care Excellence (NICE) defines febrile neutropenia (FN) or “neutropenic sepsis” as a patient with an absolute neutrophil count (ANC) < 500 cells/ml and temperature above 38 C or signs and symptoms of sepsis.2 FN is a recognised and avoidable cause of death in children receiving treatment for cancer. Patient and parent education around understanding stages of treatment where low ANC are likely, recognising early signs of sepsis and pyrexia, and seeking urgent medical attention is essential. Further to this, education of both nurses and doctors in primary and shared care oncology centres is necessary to ensure the timely evaluation and treatment of children admitted with episodes of FN. The aggressive, protocol driven management of FN with prompt nursing and medical assessment of patients, and empiric, intravenous, antimicrobial therapy has seen mortality rates reduced from approximately 30% to less than 1% toward the end of the last century.3 Monotherapy with a broad spectrum intravenous agent, either piperacillin/tazobactam or meropenem, is recommended. Following initial treatment and stabilisation of the patient, antibiotic therapy should be tailored to individual patient, guided by previous infective episodes, surveillance cultures (where available), allergies, and identified focal infection. Data on the efficacy of monotherapy vs combined agents for empirical therapy in FN has shown similar clinical outcomes and, in view of increased risk of adverse effects from combination therapy (i.e. nephrotoxicity associated with the use of aminoglycosides), monotherapy is now recommended by all international guidelines.2,4,5 An audit comparing FN practice in UK PTC demonstrated an increase in the use of piperacillin/aminoglycoside combination (7 increased to 17) and piptazobactam monotherapy (0 increased to 4) between 2005 and 2012, with a general decrease in other antibiotic therapies.6
Prevention Strategies for the prevention of FN secondary to bacteraemia and sepsis include effective hand hygiene, adherence to infection control policies, the incorporation of aseptic procedures such as Aseptic Non-Touch Techniques (ANTT) into routine practice, and patient and family
C.D. Barton et al. education. Many chemotherapy protocols include guidance and strategies to facilitate ANC recovery between the administration of subsequent course to avoid prolonged episodes of neutropaenia. Within the paediatric setting, including paediatric intensive care but also on medical and surgical wards, CVL care bundle implementations have demonstrated effectiveness in reducing the incidence of Central Lines Associated Bloodstream Infections (CLABSI).7,8 There is also recently published evidence on central venous lines (CVL) care bundle strategies being effective, albeit to a lesser extent, in paediatric oncology patients.9 The difficulty in reducing CLABSI in oncology patients may be related to factors such as the underlying immunosuppression and the loss of mucosal integrity leading to the translocation of microorganisms in the bloodstream.7,10 The 2014 CDC CLABSI definition now takes this factor into account and has added a definition for Mucosal Barrier Injury Laboratory-Confirmed Bloodstream Infection (MBI-LCBI) for oncology patients in this latest guidance.11 Further work is needed to develop strategies to further reduce CLABSI’s in paediatric oncology patients.
Risk assessment Clinical risk assessment Risk assessment involves the assessment and consideration of a child’s risk of actually developing an infection, given their underlying diagnosis and treatment. Examples of high risk patients include children with Down’s Syndrome, children receiving high dose chemotherapy as first line treatment (e.g. osteosarcoma, AML), relapse therapy (e.g. fludarabine in relapse ALL), and patients at risk of prolonged episodes of neutropaenia (e.g. patients undergoing HSTC). While there is a small proportion of patients at increased risk of sepsis, approximately 79% of episodes of febrile neutropaenia are not associated with serious infection.1 Identification of low risk patients, and out-patient management with oral antibiotics would reduce the risk of hospital acquired infection, reduce the cost of clinical care as well as the potential for development of multi-drug antibiotic resistance while having a positive impact on quality of life for patients and their families.
Risk assessment using biomarkers Research exploring biomarkers such as serum cytokines (e.g. interleukins (IL) e IL-6, IL-8, IL-10), C-reactive protein and pro-calcitonin have demonstrated potential utility for risk stratification of FN.12e15 In particular, the strategies looking at combination of several of these biomarkers have shown high sensitivity and negative predictive value (>90%) for the presence of a culture positive bacterial infection.14 However, the paucity of data on the value of these biomarkers in children, and their unavailability in most clinical laboratories means that further work is needed before they can be integrated into routine clinical practice.16
Please cite this article in press as: Barton CD, et al., Febrile neutropenia in children treated for malignancy, J Infect (2015), http:// dx.doi.org/10.1016/j.jinf.2015.04.026
63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124
YJINF3525_proof ■ 30 April 2015 ■ 3/9
Febrile neutropenia in children 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62
3
Risk prediction models (RPM) and clinical decision rules (CDR)
used clinically to identify those children at low risk at presentation with FN.
RPM takes into account multiple factors, including clinical parameters, biomarker and biochemical data, microbiological and epidemiological data, as well as demographic information to predict the risk of a future health outcome (e.g. death or adverse outcome in FN). Each CDR recommends a particular action or clinical decision based upon the prediction of an RPM (e.g. discontinue antibiotics, change to oral antibiotics). CDR’s have been developed for both adults (e.g. MASCC Risk Index17) and children.18e21 CDR such as the MASCC tool have been validated as successfully identifying low risk adult patients appropriate for out-patient management, on either IV or oral medications, with on-going review. However, this data cannot simply be extrapolated to paediatric populations, not least because of the significant changes in biology and physiology that occur between the ages of 0e18 years, but also that paediatric cancer is represented by small, heterogeneous populations of different tumour types. Paediatric research in CDR has been heterogeneous across different geographical locations and population types, with little standardisation in outcomes measured. This has proved troublesome when attempting to define a universal CDR based on the available literature. The conclusion of a recent systematic review,22 was that no individual rule is better than another, and that all CDRs must be validated locally before implementation. An institution’s choice of strategy should be influenced by:
Novel diagnostics
- Validation in the geographical location - Ability to implement rules - Availability and speed of test results. The majority of models boast moderate-high sensitivity (59e100%) with various degrees of specificity (19e100%) (Table 1) and are often over-optimistic in the discriminatory ability of the primary study compared to that at validation. Reasons for this could be different interpretations of clinical findings as well as subtle differences in the treatment regimens of specific malignancies. It has been shown in validation studies that CDR have greater discrimination 48 h after FN onset, in particular, the “Alexander rule” which is based on the absence of the following risk factors: hypotension, tachypnoea, hypoxia (SpO2 < 94%), new chest x-ray changes, altered mental state, severe mucositis, vomiting, abdominal pain or the presence of a focal infection.18,24 The use of this CDR at 48 h has shown to be safe and effective in the UK while decreasing the patient’s length of stay in the hospital and therefore is recommended by NICE as a suitable example of CDR for use in children in this setting.2 Formal randomised control trials of management strategies in low risk patients are difficult in view of the low frequency of serious or adverse events. None the less, evidence suggests that in defined settings, once patients with low risk FN have been identified, they can be rapidly stepped down to oral antibiotics safely.24 Further work is still needed to develop a universal CDR which could be
Imaging A fungal aetiology must be considered in children with persistent fever and neutropenia e over 72 h e when no source has been identified. In children at high risk for fungal complications (i.e. AML, relapse protocols, HSCT), a high resolution CT remains the modality of choice for the early identification of fungal disease, such as IPA. Early diagnosis and initiation of antifungal has shown to be important for survival in IPA and therefore should not be delayed.25 A suggestive finding for IPA on high resolution CT include the “Halo Sign” (a macrodule surrounded by a zone of ground glass appearance) which appears early in the disease process.26e28 If the CT is delayed for a few days, radiological findings are generally more non-specific in IPA, with generalised parenchymal consolidation. This eventually leads to the appearance of an “air-crescent” sign in some patients, seen upon neutrophil recovery, and a good prognostic sign (Fig. 1).29 These various radiological findings seen in adults with IPA, in particular the halo sign, have not been consistently reported in the paediatric literature, probably due to the timing of the CT performed in these studies.30,31
Galactomannan Galactomannan testing for IPA is finding increasing application as a routine investigation of suspected fungal disease in both adults and children.32 Sequential testing of plasma specimen (e.g. twice weekly) has shown to have some utility in the investigation of suspected aspergillosis in children, with a sensitivity of around 76%.33 Higher sensitivity of 78e82% has been found in analysis of single BAL specimen in children with IPA, with a specificity of 87e100%.34,35 Galactomannan testing in BAL specimen is now included in the Fourth European Conference on Infections in Leukaemia (ECIL-4) guidelines for diagnosis, prevention, and treatment of invasive fungal diseases in paediatric patients with cancer or allogeneic HSCT.36
Molecular diagnostics Increasingly, molecular techniques have been developed and validated to characterise organisms causing sepsis, with faster turnaround times and more sensitive results. A recent study using targeted RT-PCR analysis of serum of children with cancer high risk febrile neutropaenia identified pathogens in 20% of cases, compared with positive blood cultures in only 5%.37 Targeted multiplex PCR platforms identify viral (respiratory samples) and bacterial (blood specimen) pathogens on a single platform, with turnaround times of 60e90 min and sensitivities and specificities of >90%, including specimens from paediatric oncology patients.38,39 In addition to the identification of pathogens, this type of assay can
Please cite this article in press as: Barton CD, et al., Febrile neutropenia in children treated for malignancy, J Infect (2015), http:// dx.doi.org/10.1016/j.jinf.2015.04.026
Q2
63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 4
Development and validation of clinical decision rules for predicting febrile neutropenia risk in children.
RULE Validation study
Number of FN episodes
Alexander (2002)18
188
Validation study: SPOG 201020 Validation study: Dommett 200924 PINDA Santolaya (2001)19
423 762 447
Validation study: Santolaya 200252 Validation study: SPOG 201020
263 423
Validation study: Macher 201053 SPOG Ammann 201020
377 423
Miedema 201154 Amman 200355 Macher 201053 Baorto 200156 Madsen 200257 Rackoff 199658
210 227 423 377 1171 157 172
Sens. (95%CI)
Spec. (95%CI)
Comments
Absence of following RF: Hypotension Tachypnoea Hypoxia (SpO290 mg/L Hypotension Platelets
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
www.elsevierhealth.com/journals/jinf
Q8
Q7
Febrile neutropenia in children treated for malignancy Chris D. Barton a,c, Lucy K. Waugh a, Maryke J. Nielsen a, ´phane Paulus b,d,* Ste a
Department of Haematology & Oncology, Alder Hey Children’s Hospital, Eaton Road, Liverpool L12 2AP, United Kingdom b Department of Infectious Diseases, Alder Hey Children’s Hospital, Eaton Road, Liverpool L12 2AP, United Kingdom c Institute of Translational Medicine, University of Liverpool, Liverpool L69 3BX, United Kingdom d Institute of Global Health, University of Liverpool, Liverpool L69 3BX, United Kingdom Accepted 21 April 2015 Available online - - -
KEYWORDS Malignancy; Paediatrics; Febrile neutropaenia; Neutropenic sepsis
Summary Febrile Neutropenia (FN) in children treated for malignancy is a common and direct sequela of chemotherapy. Episodes of FN can be life-threatening, and demand prompt recognition, assessment and treatment with broad spectrum antibiotics. While in the majority of episodes no causal infection is identified, 10e20% are secondary to a bloodstream infection (BSI). A reduction in episodes of BSI could be achieved through robust infection prevention strategies, such as CVL care bundles. Alongside good antimicrobial stewardship, these strategies could reduce the risk of emergent, multi-drug resistant (MDR) infections. Emerging bacterial pathogens in BSI include Viridans Group Streptococci (VGS) and Enterobacteriaceae such as Klebsiella spp. which are known for their ability to carry MDR genes. There is also increased recognition of the role of invasive fungal infection (IFI) in FN, in particular with Aspergillus spp. Novel diagnostics, including multiplex blood and respiratory polymerase chain reaction assays can identify infections early in FN, facilitating targeted therapy, and reducing unnecessary antimicrobial exposure. Given appropriate, and sensitive rapid diagnostics, potential also exists to safely inform the risk assessment of patients with FN, identifying those at low risk of complication, who could be treated in the out-patient setting. Several clinical decision rules (CDR) have now been developed and validated in defined populations, for the risk assessment of children being treated for cancer. Future research is needed to develop a universal CDR to improve the management of children with FN. ª 2015 Published by Elsevier Ltd on behalf of The British Infection Association.
* Corresponding author. Alder Hey Children’s Hospital, Eaton Road, Liverpool L12 2AP, United Kingdom. Tel.: þ44 151 228 4811. E-mail addresses: [email protected] (C.D. Barton), [email protected] (L.K. Waugh), [email protected] (M.J. Nielsen), [email protected] (S. Paulus). http://dx.doi.org/10.1016/j.jinf.2015.04.026 0163-4453/ª 2015 Published by Elsevier Ltd on behalf of The British Infection Association. Please cite this article in press as: Barton CD, et al., Febrile neutropenia in children treated for malignancy, J Infect (2015), http:// dx.doi.org/10.1016/j.jinf.2015.04.026
61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122
YJINF3525_proof ■ 30 April 2015 ■ 2/9
2 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62
Introduction Children with cancer receiving cytotoxic chemotherapy and other antineoplastic therapy experience significant neutropaenia as a direct consequence of their treatment. The presence of an indwelling central line, with often concurrent loss of the mucosal integrity of the gastrointestinal tract, further predisposes these patients to bacteraemia and sepsis. Febrile episodes are observed in 34% of neutropenic periods in children treated for cancer, with bacteraemia identified in 10e20% of cases whereas an unknown aetiology is the most likely outcome (up to 79% of the cases).1
Definition and immediate management in children with febrile neutropenia The National Institute for Health and Care Excellence (NICE) defines febrile neutropenia (FN) or “neutropenic sepsis” as a patient with an absolute neutrophil count (ANC) < 500 cells/ml and temperature above 38 C or signs and symptoms of sepsis.2 FN is a recognised and avoidable cause of death in children receiving treatment for cancer. Patient and parent education around understanding stages of treatment where low ANC are likely, recognising early signs of sepsis and pyrexia, and seeking urgent medical attention is essential. Further to this, education of both nurses and doctors in primary and shared care oncology centres is necessary to ensure the timely evaluation and treatment of children admitted with episodes of FN. The aggressive, protocol driven management of FN with prompt nursing and medical assessment of patients, and empiric, intravenous, antimicrobial therapy has seen mortality rates reduced from approximately 30% to less than 1% toward the end of the last century.3 Monotherapy with a broad spectrum intravenous agent, either piperacillin/tazobactam or meropenem, is recommended. Following initial treatment and stabilisation of the patient, antibiotic therapy should be tailored to individual patient, guided by previous infective episodes, surveillance cultures (where available), allergies, and identified focal infection. Data on the efficacy of monotherapy vs combined agents for empirical therapy in FN has shown similar clinical outcomes and, in view of increased risk of adverse effects from combination therapy (i.e. nephrotoxicity associated with the use of aminoglycosides), monotherapy is now recommended by all international guidelines.2,4,5 An audit comparing FN practice in UK PTC demonstrated an increase in the use of piperacillin/aminoglycoside combination (7 increased to 17) and piptazobactam monotherapy (0 increased to 4) between 2005 and 2012, with a general decrease in other antibiotic therapies.6
Prevention Strategies for the prevention of FN secondary to bacteraemia and sepsis include effective hand hygiene, adherence to infection control policies, the incorporation of aseptic procedures such as Aseptic Non-Touch Techniques (ANTT) into routine practice, and patient and family
C.D. Barton et al. education. Many chemotherapy protocols include guidance and strategies to facilitate ANC recovery between the administration of subsequent course to avoid prolonged episodes of neutropaenia. Within the paediatric setting, including paediatric intensive care but also on medical and surgical wards, CVL care bundle implementations have demonstrated effectiveness in reducing the incidence of Central Lines Associated Bloodstream Infections (CLABSI).7,8 There is also recently published evidence on central venous lines (CVL) care bundle strategies being effective, albeit to a lesser extent, in paediatric oncology patients.9 The difficulty in reducing CLABSI in oncology patients may be related to factors such as the underlying immunosuppression and the loss of mucosal integrity leading to the translocation of microorganisms in the bloodstream.7,10 The 2014 CDC CLABSI definition now takes this factor into account and has added a definition for Mucosal Barrier Injury Laboratory-Confirmed Bloodstream Infection (MBI-LCBI) for oncology patients in this latest guidance.11 Further work is needed to develop strategies to further reduce CLABSI’s in paediatric oncology patients.
Risk assessment Clinical risk assessment Risk assessment involves the assessment and consideration of a child’s risk of actually developing an infection, given their underlying diagnosis and treatment. Examples of high risk patients include children with Down’s Syndrome, children receiving high dose chemotherapy as first line treatment (e.g. osteosarcoma, AML), relapse therapy (e.g. fludarabine in relapse ALL), and patients at risk of prolonged episodes of neutropaenia (e.g. patients undergoing HSTC). While there is a small proportion of patients at increased risk of sepsis, approximately 79% of episodes of febrile neutropaenia are not associated with serious infection.1 Identification of low risk patients, and out-patient management with oral antibiotics would reduce the risk of hospital acquired infection, reduce the cost of clinical care as well as the potential for development of multi-drug antibiotic resistance while having a positive impact on quality of life for patients and their families.
Risk assessment using biomarkers Research exploring biomarkers such as serum cytokines (e.g. interleukins (IL) e IL-6, IL-8, IL-10), C-reactive protein and pro-calcitonin have demonstrated potential utility for risk stratification of FN.12e15 In particular, the strategies looking at combination of several of these biomarkers have shown high sensitivity and negative predictive value (>90%) for the presence of a culture positive bacterial infection.14 However, the paucity of data on the value of these biomarkers in children, and their unavailability in most clinical laboratories means that further work is needed before they can be integrated into routine clinical practice.16
Please cite this article in press as: Barton CD, et al., Febrile neutropenia in children treated for malignancy, J Infect (2015), http:// dx.doi.org/10.1016/j.jinf.2015.04.026
63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124
YJINF3525_proof ■ 30 April 2015 ■ 3/9
Febrile neutropenia in children 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62
3
Risk prediction models (RPM) and clinical decision rules (CDR)
used clinically to identify those children at low risk at presentation with FN.
RPM takes into account multiple factors, including clinical parameters, biomarker and biochemical data, microbiological and epidemiological data, as well as demographic information to predict the risk of a future health outcome (e.g. death or adverse outcome in FN). Each CDR recommends a particular action or clinical decision based upon the prediction of an RPM (e.g. discontinue antibiotics, change to oral antibiotics). CDR’s have been developed for both adults (e.g. MASCC Risk Index17) and children.18e21 CDR such as the MASCC tool have been validated as successfully identifying low risk adult patients appropriate for out-patient management, on either IV or oral medications, with on-going review. However, this data cannot simply be extrapolated to paediatric populations, not least because of the significant changes in biology and physiology that occur between the ages of 0e18 years, but also that paediatric cancer is represented by small, heterogeneous populations of different tumour types. Paediatric research in CDR has been heterogeneous across different geographical locations and population types, with little standardisation in outcomes measured. This has proved troublesome when attempting to define a universal CDR based on the available literature. The conclusion of a recent systematic review,22 was that no individual rule is better than another, and that all CDRs must be validated locally before implementation. An institution’s choice of strategy should be influenced by:
Novel diagnostics
- Validation in the geographical location - Ability to implement rules - Availability and speed of test results. The majority of models boast moderate-high sensitivity (59e100%) with various degrees of specificity (19e100%) (Table 1) and are often over-optimistic in the discriminatory ability of the primary study compared to that at validation. Reasons for this could be different interpretations of clinical findings as well as subtle differences in the treatment regimens of specific malignancies. It has been shown in validation studies that CDR have greater discrimination 48 h after FN onset, in particular, the “Alexander rule” which is based on the absence of the following risk factors: hypotension, tachypnoea, hypoxia (SpO2 < 94%), new chest x-ray changes, altered mental state, severe mucositis, vomiting, abdominal pain or the presence of a focal infection.18,24 The use of this CDR at 48 h has shown to be safe and effective in the UK while decreasing the patient’s length of stay in the hospital and therefore is recommended by NICE as a suitable example of CDR for use in children in this setting.2 Formal randomised control trials of management strategies in low risk patients are difficult in view of the low frequency of serious or adverse events. None the less, evidence suggests that in defined settings, once patients with low risk FN have been identified, they can be rapidly stepped down to oral antibiotics safely.24 Further work is still needed to develop a universal CDR which could be
Imaging A fungal aetiology must be considered in children with persistent fever and neutropenia e over 72 h e when no source has been identified. In children at high risk for fungal complications (i.e. AML, relapse protocols, HSCT), a high resolution CT remains the modality of choice for the early identification of fungal disease, such as IPA. Early diagnosis and initiation of antifungal has shown to be important for survival in IPA and therefore should not be delayed.25 A suggestive finding for IPA on high resolution CT include the “Halo Sign” (a macrodule surrounded by a zone of ground glass appearance) which appears early in the disease process.26e28 If the CT is delayed for a few days, radiological findings are generally more non-specific in IPA, with generalised parenchymal consolidation. This eventually leads to the appearance of an “air-crescent” sign in some patients, seen upon neutrophil recovery, and a good prognostic sign (Fig. 1).29 These various radiological findings seen in adults with IPA, in particular the halo sign, have not been consistently reported in the paediatric literature, probably due to the timing of the CT performed in these studies.30,31
Galactomannan Galactomannan testing for IPA is finding increasing application as a routine investigation of suspected fungal disease in both adults and children.32 Sequential testing of plasma specimen (e.g. twice weekly) has shown to have some utility in the investigation of suspected aspergillosis in children, with a sensitivity of around 76%.33 Higher sensitivity of 78e82% has been found in analysis of single BAL specimen in children with IPA, with a specificity of 87e100%.34,35 Galactomannan testing in BAL specimen is now included in the Fourth European Conference on Infections in Leukaemia (ECIL-4) guidelines for diagnosis, prevention, and treatment of invasive fungal diseases in paediatric patients with cancer or allogeneic HSCT.36
Molecular diagnostics Increasingly, molecular techniques have been developed and validated to characterise organisms causing sepsis, with faster turnaround times and more sensitive results. A recent study using targeted RT-PCR analysis of serum of children with cancer high risk febrile neutropaenia identified pathogens in 20% of cases, compared with positive blood cultures in only 5%.37 Targeted multiplex PCR platforms identify viral (respiratory samples) and bacterial (blood specimen) pathogens on a single platform, with turnaround times of 60e90 min and sensitivities and specificities of >90%, including specimens from paediatric oncology patients.38,39 In addition to the identification of pathogens, this type of assay can
Please cite this article in press as: Barton CD, et al., Febrile neutropenia in children treated for malignancy, J Infect (2015), http:// dx.doi.org/10.1016/j.jinf.2015.04.026
Q2
63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 4
Development and validation of clinical decision rules for predicting febrile neutropenia risk in children.
RULE Validation study
Number of FN episodes
Alexander (2002)18
188
Validation study: SPOG 201020 Validation study: Dommett 200924 PINDA Santolaya (2001)19
423 762 447
Validation study: Santolaya 200252 Validation study: SPOG 201020
263 423
Validation study: Macher 201053 SPOG Ammann 201020
377 423
Miedema 201154 Amman 200355 Macher 201053 Baorto 200156 Madsen 200257 Rackoff 199658
210 227 423 377 1171 157 172
Sens. (95%CI)
Spec. (95%CI)
Comments
Absence of following RF: Hypotension Tachypnoea Hypoxia (SpO290 mg/L Hypotension Platelets
