Original Studies
Lack of Accuracy of Body Temperature for Detecting Serious Bacterial Infection in Febrile Episodes Sukanya De, MB BS, MD, PhD, FRACP,*† Gabrielle J. Williams, BSc, MPH, PhD, *† Armando Teixeira-Pinto, BSc, AM, PhD,* Petra Macaskill, BA, MApp Stat, PhD,* Mary McCaskill, BSc, Dip Paeds, MB BS, FACEM,‡ David Isaacs, MB BChir, MD, FRACP, FRCPCH,§ and Jonathan C. Craig, MB ChB, DipCH, MM, PhD, FRACP*†¶ Background: Body temperature is a time-honored marker of serious bacterial infection, but there are few studies of its test performance. The aim of our study was to determine the accuracy of temperature measured on presentation to medical care for detecting serious bacterial infection. Methods: Febrile children 0–5 years of age presenting to the emergency department of a tertiary care pediatric hospital were sampled consecutively. The accuracy of the axillary temperature measured at presentation was evaluated using logistic regression models to generate receiver operating characteristic curves. Reference standard tests for serious bacterial infection were standard microbiologic/radiologic tests and clinical follow-up. Age, clinicians’ impression of appearance of the child (well versus unwell) and duration of illness were assessed as possible effect modifiers. Results: Of 15,781 illness episodes 1120 (7.1%) had serious bacterial infection. The area under the receiver operating characteristic curve for temperature was 0.60 [95% confidence intervals (CI): 0.58–0.62]. A threshold of ≥38°C had a sensitivity of 0.67 (95% CI: 0.64–0.70), specificity of 0.45 (95% CI: 0.44– 0.46), positive likelihood ratio of 1.2 (95% CI: 1.2–1.3) and negative likelihood ratio of 0.7 (95% CI: 0.7–0.8). Age and illness duration had a small but significant effect on the accuracy of temperature increasing its “rule-in” potential. Conclusion: Measured temperature at presentation to hospital is not an accurate marker of serious bacterial infection in febrile children. Younger age and longer duration of illness increase the rule-in potential of temperature but without substantial overall change in its test accuracy. Key Words: fever, sensitivity and specificity, serious bacterial infection (Pediatr Infect Dis J 2015;34:940–944)
self-resolving viral infections, depending on the setting, between 7% and 27% are because of a serious bacterial infection.3–5 Clinical detection of serious bacterial infections in young children is challenging as many will present without localizing signs.6 The National Institute for Health and Care Excellence guideline advises that although height of fever alone should not be used to identify children with serious illness, a temperature over 39°C in 3–6 months old infants warrants a heightened level of concern and consideration of further tests.7 The American College of Emergency Physicians’ clinical policy for young children presenting to emergency department with fever recommends consideration of empiric antibiotics therapy for previously healthy, well-appearing children between 3 and 36 months of age with fever without a source, a body temperature above 39°C and white blood cell count of 15 × 109/L or greater.8 Body temperature has been included in several clinical decision rules for determining likelihood of serious bacterial infection.9–11 The primary aim of this study was to evaluate whether elevated body temperature at presentation to hospital is an accurate clinical marker of serious bacterial infection in young febrile children and therefore a useful guide for clinicians to determine the need for further tests and treatment. Our secondary aim was to determine whether the accuracy of temperature was constant across all groups of children or whether it was better in children with a prolonged duration of illness, a younger age or in children who appeared unwell.
METHODS
F
ebrile illnesses are common in young children, and fever accounts for up to a third of presentations to pediatric emergency departments.1 Body temperature outside of the normal range generates considerable parental anxiety prompting them to seek medical advice.2 Although most febrile episodes are caused by Accepted for publication April 20, 2015 From the *Screening and Test Evaluation Program, Sydney School of Public Health, The University of Sydney, Sydney, Australia; †Centre for Kidney Research, ‡Department of Emergency Medicine, §Department of Infectious Disease, and ¶Department of Nephrology, The Children’s Hospital at Westmead, Sydney, Australia. This is a substudy of the Febrile Evaluation of Children in the Emergency Room (FEVER) study, which was funded by the National Health and Medical Research Council of Australia (program grant numbers 211205 and 402764). The funding source had no influence on study design, data collection, analysis, interpretation of data, writing of the report, or on the decision to submit the paper for publication. The authors have no conflicts of interest to disclose. Address for correspondence: Sukanya De, Centre for Kidney Research, The Children’s Hospital at Westmead, Westmead, NSW 2145, Australia. E-mail: [email protected]. Supplemental digital content is available for this article. Direct URL citations appear in the printed text and are provided in the HTML and PDF versions of this article on the journal’s website (www.pidj.com). Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved. ISSN: 0891-3668/15/3409-0940 DOI: 10.1097/INF.0000000000000771
940 | www.pidj.com
We applied the Standard for Reporting of Diagnostic Accuracy guidelines for study reporting purposes.12
Study Design and Setting This is a substudy of the Febrile Evaluation of Children in Emergency Department (FEVER), which was a prospective cohort study, conducted in the Emergency Department of the Children’s Hospital at Westmead between July 1, 2004 and June 30, 2006. Detailed methods have been reported elsewhere.3
Recruitment During the study period, consecutive sampling of febrile illnesses (in children aged less than 5 years) presenting to the emergency department was done. Febrile illness was defined as any illness that met one or more of the following criteria: a measured axillary temperature of ≥38.0°C; parental report of a temperature of ≥38.0°C measured at home within the previous 24 hours and parental report that the child “ felt hot” in the previous 24 hours.
Study Exclusion Criteria Children transferred from another hospital, those with malignancy and transplant recipients, were excluded. The unit of analysis was an instance of febrile illness. In the case of multiple presentations with the same illness, history and clinical evaluation
The Pediatric Infectious Disease Journal • Volume 34, Number 9, September 2015
Copyright © 2015 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
The Pediatric Infectious Disease Journal • Volume 34, Number 9, September 2015
data from the first visit only were used. Case definition for “same illness” was if the child presented within 24 hours of a previous visit or if the fever had persisted between visits to the emergency department without a fever free period of at least 24 hours.
Data Collection Clinical information for each illness episode were entered by the examining physician (pediatric emergency physician, pediatric or emergency medicine trainee) into a mandatory febrile child assessment template, added to the hospital’s electronic record keeping system (details of the items in this template are provided elsewhere).3 The template was completed following the initial physician assessment but before test results were obtained. This information was entered into a research data base and included a comprehensive list of symptoms, findings on examination including examining physician’s impression of appearance of the child (well, mildly unwell or very unwell), duration of the illness and significant background medical information. Records of investigations carried out at the hospital were electronically linked by unique medical record number and date to within 3 days of triage date, of eligible febrile visits.
Index Test Definition The index test was the highest axillary temperature measurement recorded on presentation to the emergency department. This was done by the triage nurse, using an axillary digital thermometer (Terumo ET*C205S). Repeat measurements were undertaken during the waiting period before the physicians’ assessment at the triage nurses discretion. The number of measurements varied per child, ranging from 1 to 5 measurements, although most children (71%) had only 1 measurement. During the initial clinical assessment, a temperature reading was also recorded by the examining physician. For this study, the highest of the temperature reading for each child, among those recorded by the triage nurse and during initial assessment by the examining physician, was used as the index test. Because the temperature recorded at presentation to hospital may not capture the maximum temperature experienced during the illness and to gauge whether temperature observed over a more extended period of time was a better marker, we also evaluated the accuracy of the maximum temperature in the previous 24 hours as reported by carers to (1) the triage nurse and (2) the examining physician. At presentation, a triage nurse asked each carer; “Has your child’s temperature been taken in the past 24 hours?” and if yes, “What was the highest recorded temperature in the past 24 hours?” according to the 4 categories: ≤37.9, 38–38.9, 39–39.9 and ≥40°C. These questions were components of a mandatory triage process. During physician assessment of a febrile child, similar questions were asked by the doctor and recorded in the structured febrile child assessment template.
Reference Standard Test and Outcome Classification Diagnoses of serious bacterial infections were based on criteria for culture positivity of usually sterile fluid or radiological criteria (for pneumonia) and have been detailed elsewhere.3 In the FEVER study, the decision to perform confirmatory tests for identification of bacterial infections was at the treating physicians’ discretion because mandatory testing of blood, urine and other body fluids of all febrile children is not standard clinical practice and could not be ethically justifiable. To ensure correct outcome classification, we applied a double reference standard for determining infection status. This composed of microbiological/radiological confirmation of serious bacterial infection if performed and © 2015 Wolters Kluwer Health, Inc. All rights reserved.
Body Temperature and Infection
follow-up of all enrolled study participants. All probable cases of serious bacterial infection were reviewed by an expert committee composed of 2 specialist pediatricians (with expertise in pediatrics, infectious disease and respiratory medicine) and in cases of pneumonia a radiologist, who reviewed test results blinded to clinical information. All study participants were followed-up until they either fulfilled the case definition for serious bacterial infection or until the fever had resolved for over 24 hours. Follow-up included review of hospital records including test details, duration of admission, representations and parental phone contact at days 10–14. Results of tests done at any other health care facility were sought after obtaining parental permission. For this study, we classified children as either having a serious bacterial infection or not having a serious bacterial infection. The latter category included children with clinically diagnosed infection (where a clinical test was positive and it was possible that the infection was of bacterial origin such as otitis media, tonsillitis, cellulitis or abscess) and children with no bacterial infection (no serious bacterial infection or clinically diagnosed infection and spontaneous illness resolution by days 10–14).
Statistical Analysis Sensitivity and specificity were computed at 3 thresholds (38, 39 and 40°C) of the maximum temperature recorded at presentation, and exact 95% confidence intervals (CI) were computed for these estimates. The maximum temperature at presentation was recorded as a continuous measurement. We evaluated the overall test accuracy of the temperature recorded upon presentation to hospital as a marker of serious bacterial infection using receiver operating characteristics (ROC) curves and the respective areas under the curves (AUC). We generated plots displaying the distribution of temperature on presentation to the emergency department (grouped according to the presence or absence of serious bacterial infection) using a Gaussian kernel for smoothing. We also evaluated the test accuracy of the maximum temperature in the previous 24 hours as reported by care givers to the triage nurse and the examining physician (recorded on an ordinal scale) using ROC curve analysis. The weighted kappa statistic with linear weights was used to evaluate agreement between the temperature reported to the nurse and the physician. Separate logistic regression models were used to test for interaction between temperature at presentation and (i) age of the child (0 to
Lack of Accuracy of Body Temperature for Detecting Serious Bacterial Infection in Febrile Episodes Sukanya De, MB BS, MD, PhD, FRACP,*† Gabrielle J. Williams, BSc, MPH, PhD, *† Armando Teixeira-Pinto, BSc, AM, PhD,* Petra Macaskill, BA, MApp Stat, PhD,* Mary McCaskill, BSc, Dip Paeds, MB BS, FACEM,‡ David Isaacs, MB BChir, MD, FRACP, FRCPCH,§ and Jonathan C. Craig, MB ChB, DipCH, MM, PhD, FRACP*†¶ Background: Body temperature is a time-honored marker of serious bacterial infection, but there are few studies of its test performance. The aim of our study was to determine the accuracy of temperature measured on presentation to medical care for detecting serious bacterial infection. Methods: Febrile children 0–5 years of age presenting to the emergency department of a tertiary care pediatric hospital were sampled consecutively. The accuracy of the axillary temperature measured at presentation was evaluated using logistic regression models to generate receiver operating characteristic curves. Reference standard tests for serious bacterial infection were standard microbiologic/radiologic tests and clinical follow-up. Age, clinicians’ impression of appearance of the child (well versus unwell) and duration of illness were assessed as possible effect modifiers. Results: Of 15,781 illness episodes 1120 (7.1%) had serious bacterial infection. The area under the receiver operating characteristic curve for temperature was 0.60 [95% confidence intervals (CI): 0.58–0.62]. A threshold of ≥38°C had a sensitivity of 0.67 (95% CI: 0.64–0.70), specificity of 0.45 (95% CI: 0.44– 0.46), positive likelihood ratio of 1.2 (95% CI: 1.2–1.3) and negative likelihood ratio of 0.7 (95% CI: 0.7–0.8). Age and illness duration had a small but significant effect on the accuracy of temperature increasing its “rule-in” potential. Conclusion: Measured temperature at presentation to hospital is not an accurate marker of serious bacterial infection in febrile children. Younger age and longer duration of illness increase the rule-in potential of temperature but without substantial overall change in its test accuracy. Key Words: fever, sensitivity and specificity, serious bacterial infection (Pediatr Infect Dis J 2015;34:940–944)
self-resolving viral infections, depending on the setting, between 7% and 27% are because of a serious bacterial infection.3–5 Clinical detection of serious bacterial infections in young children is challenging as many will present without localizing signs.6 The National Institute for Health and Care Excellence guideline advises that although height of fever alone should not be used to identify children with serious illness, a temperature over 39°C in 3–6 months old infants warrants a heightened level of concern and consideration of further tests.7 The American College of Emergency Physicians’ clinical policy for young children presenting to emergency department with fever recommends consideration of empiric antibiotics therapy for previously healthy, well-appearing children between 3 and 36 months of age with fever without a source, a body temperature above 39°C and white blood cell count of 15 × 109/L or greater.8 Body temperature has been included in several clinical decision rules for determining likelihood of serious bacterial infection.9–11 The primary aim of this study was to evaluate whether elevated body temperature at presentation to hospital is an accurate clinical marker of serious bacterial infection in young febrile children and therefore a useful guide for clinicians to determine the need for further tests and treatment. Our secondary aim was to determine whether the accuracy of temperature was constant across all groups of children or whether it was better in children with a prolonged duration of illness, a younger age or in children who appeared unwell.
METHODS
F
ebrile illnesses are common in young children, and fever accounts for up to a third of presentations to pediatric emergency departments.1 Body temperature outside of the normal range generates considerable parental anxiety prompting them to seek medical advice.2 Although most febrile episodes are caused by Accepted for publication April 20, 2015 From the *Screening and Test Evaluation Program, Sydney School of Public Health, The University of Sydney, Sydney, Australia; †Centre for Kidney Research, ‡Department of Emergency Medicine, §Department of Infectious Disease, and ¶Department of Nephrology, The Children’s Hospital at Westmead, Sydney, Australia. This is a substudy of the Febrile Evaluation of Children in the Emergency Room (FEVER) study, which was funded by the National Health and Medical Research Council of Australia (program grant numbers 211205 and 402764). The funding source had no influence on study design, data collection, analysis, interpretation of data, writing of the report, or on the decision to submit the paper for publication. The authors have no conflicts of interest to disclose. Address for correspondence: Sukanya De, Centre for Kidney Research, The Children’s Hospital at Westmead, Westmead, NSW 2145, Australia. E-mail: [email protected]. Supplemental digital content is available for this article. Direct URL citations appear in the printed text and are provided in the HTML and PDF versions of this article on the journal’s website (www.pidj.com). Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved. ISSN: 0891-3668/15/3409-0940 DOI: 10.1097/INF.0000000000000771
940 | www.pidj.com
We applied the Standard for Reporting of Diagnostic Accuracy guidelines for study reporting purposes.12
Study Design and Setting This is a substudy of the Febrile Evaluation of Children in Emergency Department (FEVER), which was a prospective cohort study, conducted in the Emergency Department of the Children’s Hospital at Westmead between July 1, 2004 and June 30, 2006. Detailed methods have been reported elsewhere.3
Recruitment During the study period, consecutive sampling of febrile illnesses (in children aged less than 5 years) presenting to the emergency department was done. Febrile illness was defined as any illness that met one or more of the following criteria: a measured axillary temperature of ≥38.0°C; parental report of a temperature of ≥38.0°C measured at home within the previous 24 hours and parental report that the child “ felt hot” in the previous 24 hours.
Study Exclusion Criteria Children transferred from another hospital, those with malignancy and transplant recipients, were excluded. The unit of analysis was an instance of febrile illness. In the case of multiple presentations with the same illness, history and clinical evaluation
The Pediatric Infectious Disease Journal • Volume 34, Number 9, September 2015
Copyright © 2015 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
The Pediatric Infectious Disease Journal • Volume 34, Number 9, September 2015
data from the first visit only were used. Case definition for “same illness” was if the child presented within 24 hours of a previous visit or if the fever had persisted between visits to the emergency department without a fever free period of at least 24 hours.
Data Collection Clinical information for each illness episode were entered by the examining physician (pediatric emergency physician, pediatric or emergency medicine trainee) into a mandatory febrile child assessment template, added to the hospital’s electronic record keeping system (details of the items in this template are provided elsewhere).3 The template was completed following the initial physician assessment but before test results were obtained. This information was entered into a research data base and included a comprehensive list of symptoms, findings on examination including examining physician’s impression of appearance of the child (well, mildly unwell or very unwell), duration of the illness and significant background medical information. Records of investigations carried out at the hospital were electronically linked by unique medical record number and date to within 3 days of triage date, of eligible febrile visits.
Index Test Definition The index test was the highest axillary temperature measurement recorded on presentation to the emergency department. This was done by the triage nurse, using an axillary digital thermometer (Terumo ET*C205S). Repeat measurements were undertaken during the waiting period before the physicians’ assessment at the triage nurses discretion. The number of measurements varied per child, ranging from 1 to 5 measurements, although most children (71%) had only 1 measurement. During the initial clinical assessment, a temperature reading was also recorded by the examining physician. For this study, the highest of the temperature reading for each child, among those recorded by the triage nurse and during initial assessment by the examining physician, was used as the index test. Because the temperature recorded at presentation to hospital may not capture the maximum temperature experienced during the illness and to gauge whether temperature observed over a more extended period of time was a better marker, we also evaluated the accuracy of the maximum temperature in the previous 24 hours as reported by carers to (1) the triage nurse and (2) the examining physician. At presentation, a triage nurse asked each carer; “Has your child’s temperature been taken in the past 24 hours?” and if yes, “What was the highest recorded temperature in the past 24 hours?” according to the 4 categories: ≤37.9, 38–38.9, 39–39.9 and ≥40°C. These questions were components of a mandatory triage process. During physician assessment of a febrile child, similar questions were asked by the doctor and recorded in the structured febrile child assessment template.
Reference Standard Test and Outcome Classification Diagnoses of serious bacterial infections were based on criteria for culture positivity of usually sterile fluid or radiological criteria (for pneumonia) and have been detailed elsewhere.3 In the FEVER study, the decision to perform confirmatory tests for identification of bacterial infections was at the treating physicians’ discretion because mandatory testing of blood, urine and other body fluids of all febrile children is not standard clinical practice and could not be ethically justifiable. To ensure correct outcome classification, we applied a double reference standard for determining infection status. This composed of microbiological/radiological confirmation of serious bacterial infection if performed and © 2015 Wolters Kluwer Health, Inc. All rights reserved.
Body Temperature and Infection
follow-up of all enrolled study participants. All probable cases of serious bacterial infection were reviewed by an expert committee composed of 2 specialist pediatricians (with expertise in pediatrics, infectious disease and respiratory medicine) and in cases of pneumonia a radiologist, who reviewed test results blinded to clinical information. All study participants were followed-up until they either fulfilled the case definition for serious bacterial infection or until the fever had resolved for over 24 hours. Follow-up included review of hospital records including test details, duration of admission, representations and parental phone contact at days 10–14. Results of tests done at any other health care facility were sought after obtaining parental permission. For this study, we classified children as either having a serious bacterial infection or not having a serious bacterial infection. The latter category included children with clinically diagnosed infection (where a clinical test was positive and it was possible that the infection was of bacterial origin such as otitis media, tonsillitis, cellulitis or abscess) and children with no bacterial infection (no serious bacterial infection or clinically diagnosed infection and spontaneous illness resolution by days 10–14).
Statistical Analysis Sensitivity and specificity were computed at 3 thresholds (38, 39 and 40°C) of the maximum temperature recorded at presentation, and exact 95% confidence intervals (CI) were computed for these estimates. The maximum temperature at presentation was recorded as a continuous measurement. We evaluated the overall test accuracy of the temperature recorded upon presentation to hospital as a marker of serious bacterial infection using receiver operating characteristics (ROC) curves and the respective areas under the curves (AUC). We generated plots displaying the distribution of temperature on presentation to the emergency department (grouped according to the presence or absence of serious bacterial infection) using a Gaussian kernel for smoothing. We also evaluated the test accuracy of the maximum temperature in the previous 24 hours as reported by care givers to the triage nurse and the examining physician (recorded on an ordinal scale) using ROC curve analysis. The weighted kappa statistic with linear weights was used to evaluate agreement between the temperature reported to the nurse and the physician. Separate logistic regression models were used to test for interaction between temperature at presentation and (i) age of the child (0 to
