Timing of antibiotics in febrile children meeting sepsis criteria at a paediatric emergency department

Delay in antibiotic administration in paediatric sepsis is associated with increased mortality and prolonged organ dysfunction. This pre‐intervention study evaluated performance in paediatric sepsis management.


Introduction
2][3] A retrospective study in Australia and New Zealand showed that intensive care unit (ICU) case fatality due to sepsis and septic shock was much higher among paediatric admissions. 46][7][8] New South Wales (NSW) Clinical Excellence Commission (CEC) Paediatric Sepsis Pathway defines sepsis as infection, either suspected or confirmed, with signs of toxicity and systemic inflammatory response syndrome (SIRS). 9Signs of systemic illness can be present in febrile children with or without infection leading to low specificity and poor positive predictive value of this definition. 10The international consensus on sepsis in 2016 moved away from SIRS in favour of a Sequential Organ Failure Assessment (SOFA) score, but this specifically excluded the paediatric population. 11A review of a quick SOFA (qSOFA) score in children in EDs, including mentation, systolic blood pressure and respiratory rate, has shown moderate prognostic accuracy for intensive care admission or mortality. 12,13Notwithstanding these challenges, sepsis recognition is critical to improving clinical outcomes.
The key determinants for the outcome for a septic patient are timely recognition, resuscitation and This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited.

Accepted 2 July 2023
Key findings stabilisation that includes timely administration of antibiotic therapy. 14Delay in antibiotic administration is associated with increased mortality and prolonged organ dysfunction.Despite this, it has been found that antibiotics are not always administered within recommended timeframes. 8,15,16he 2014 American College of Critical Care Medicine guideline emphasises the importance of implementation of locally developed sepsis bundles that include Recognition, Resuscitation, Stabilisation and Performance. 17The Surviving Sepsis Campaign 2020 (SSCM 2020) guidelines reaffirm these recommendations and provide further guidance on the timing of interventions including antimicrobial therapy. 8Antibiotics should be started within 60 min in septic shock or 3 h in sepsis-associated organ dysfunction without shock.Recently, alerts on electronic medical record platforms have been shown to lead to improved recognition of sepsis. 18he CEC introduced the Sepsis Kills programme in 2011.This aims to improve the recognition and treatment of sepsis by outlining the signs and symptoms of paediatric sepsis and recommending the administration of antibiotics within 60 min when this is suspected.
In the ED at The Children's Hospital at Westmead, delay in the administration of intravenous antibiotics for children with sepsis was identified as a major concern.In the present study, we aimed to determine the frequency and factors leading to delay in the administration of antibiotics for children meeting CEC Paediatric Sepsis criteria.

Design and setting
We conducted a retrospective study of children admitted with fever through a tertiary paediatric ED with over 55 000 presentations annually.This was followed by an online survey of ED clinicians regarding the management of children with sepsis.
All ED admission encounters between 1 May and 31 July 2017 with fever were enrolled.We manually reviewed clinical documentation and extracted data from PowerChart (Cerner ® ) electronic medical records.We included children aged 29 days to 60 months with fever (axillary temperature ≥38 C) who were admitted through ED.We excluded neonates (established neonatal sepsis protocols in place), simple febrile seizures and children who received intravenous antibiotics before arrival in the ED.Enrolment procedures are presented in Figure 1.
We searched for clinical information in the admission encounter.This included CEC Paediatric Sepsis Pathway risk factors for sepsis, signs of toxicity and systemic inflammation. 9isk factors for sepsis included representation within 48 h, deterioration despite treatment, 90 days or younger, a high level of parental concern, immunocompromising medical conditions or treatment, recent surgery or traumatic wounds, a central venous access device or other invasive devices in situ.Signs of toxicity include decreased alertness, arousal or activity, pale or mottled colour, cool peripheries, weak cry, grunting, rigours and bounding pulses.Age-stratified NSW standard paediatric observation charts were used to assess heart rate, respiratory rate and respiratory effort at presentation as moderately abnormal (yellow zone) or severely abnormal (red zone). 19Children were considered to have signs of systemic inflammation if they had tachycardia (heart rate >95th centile for age) and tachypnoea (respiratory rate >95th centile for age) with or without leukocytosis, leukopenia, or bandaemia (white blood cells >12 000/mm 3 , <4000/mm 3 or bandaemia ≥10%).We assessed alignment of choice of antibiotic therapy with CEC antibiotic guidelines recommendations.
We applied the CEC Paediatric Sepsis Pathway 9 to clinical information in the patient records.Sepsis criteria were met if a patient with a suspected or confirmed infection had all the following: fever (axillary temperature >38 C), signs of toxicity, tachycardia or tachypnoea with or without or abnormal white cell count (Fig. 1).The diagnosis was documented according to the International Classification of Diseases (ICD-10) codes.
A serious bacterial infection was defined as bacteraemia, bacterial meningitis, urinary tract infection (UTI) or pneumonia.Severe sepsis was defined as sepsis plus cardiovascular organ dysfunction or acute respiratory distress syndrome or two or more other organ dysfunctions.Septic shock was defined as sepsis with sepsis-induced cardiovascular organ dysfunction.The primary outcome was to determine the frequency of delay in the administration of intravenous antibiotics in children meeting CEC sepsis criteria.The baseline time for all interventions was triage.Delay was defined as time exceeding 60 min from triage to complete administration of the first dose of intravenous antibiotics in suspected sepsis.Secondary outcomes were to determine the proportion of children meeting sepsis criteria treated with intravenous antibiotics, identify common reasons for delay and factors influencing timing of antibiotic administration.
We further compared outcomes for two age categories of infants aged <1 year and preschool-aged children 1-5 years as well as immunocompromised children and children with central venous access device in situ.
To supplement the limited documentation of clinical decisions leading to delay of intravenous antibiotics, we conducted an online survey of ED medical doctors regarding paediatric sepsis management.The survey was posted on SurveyMonkey ® between July and August 2018.Survey questions were identified from previous departmental clinical meetings including morbidity and mortality meetings.

Statistical analysis
Data analysis was conducted using STATA 2014 (College Station, TX, USA, www.stata.com/stata14).The data were analysed for the whole group and then for infants <1 year and preschool-aged children 1-5 years.The mean and standard deviation (SD) were calculated for normally distributed data and the median and interquartile range (IQR) otherwise.Analysis for statistical significance was carried out using parametric methods for normally distributed data.The chisquared test was used for categorical variables.We conducted a t-test for independent samples when comparing two samples and analysis of variance to compare three or more independent samples.We conducted two-sample Wilcoxon rank-sum test for non-normally distributed continuous variables.Significance was set with an alpha 0.05.

Ethics
The study plan was submitted as a quality improvement activity and received approval from the institutional ethics committee.Ethics approval number QIE 2011-10-15.
The mean Socio-Economic Indexes for Areas rank was 1372 (SD 846) and median Socio-Economic Indexes for Areas decile of 6 (IQR 2-9) with no difference between age groups or those meeting sepsis criteria.The median length of stay in ED was approximately 7 h.

Clinical presentation
Clinical features at presentation are presented in Table 1.Children 2-12 months were more likely to have signs of systemic inflammation and meet sepsis criteria (P < 0.01).Severe tachycardia and tachypnoea at triage were recorded in 44 (31%) and eight (6%) children, respectively.Six (4%) children had a clinician diagnosis of sepsis on admission (one did not meet local sepsis criteria), four (3%) of whom had a diagnosis of sepsis at discharge.Thirty-seven (26%) febrile children met local sepsis criteria.
Two (6%) children were admitted to ICU due to other diagnoses.One (3%) child had acute adrenal insufficiency, the other child with Dravet's syndrome had febrile status epilepticus.No children received vasopressor therapy.No children had septic shock.No children died during admission.

Antibiotic administration
Timing of antibiotic administration is presented in Figure 2. Time to intravenous access was documented in 34 (24%) out of 144 children, whereas time to antibiotics was recorded in all patients who received antibiotics.Delay in delivery of antibiotics occurred in 26 (76%) children meeting sepsis criteria (Table 1).However, children meeting sepsis criteria received antibiotics earlier than their counterparts, although this was not statistically significant.In children meeting sepsis criteria, eight (22%) children received intravenous antibiotics within 60 min and 25 (68%) children within 180 min.Three children meeting sepsis criteria did not receive antibiotics.Among the six children with an admission diagnosis of sepsis, half received antibiotics within 60 min and rest after 180 min.The choice of antibiotics followed the appropriate CEC guidelines in all instances. 12easons for delay of antibiotic therapy were documented in 12 (32%) medical charts.These included clinicians opting for clinical observation and reassessment in four (11%), awaiting investigation results to guide therapeutic decisions in three (8%), child not considered septic in two (5%), no vascular access in one (3%), other two (6%).

Fluid resuscitation therapy
Fluid resuscitation therapy of 10 or 20 mL/kg of 0.9% sodium chloride intravenous boluses were given to dehydrated or poorly perfused children.Children meeting sepsis criteria had fluid resuscitation therapy more frequently (P < 0.001).

Immunocompromised children
We identified 41 (39%) immunocompromised children, 21 with central venous access devices at presentation.Eight of these children met sepsis criteria at presentation; two had positive blood cultures.One child had a discharge diagnosis of sepsis.The median time to antibiotics in this group was 69 min (IQR 48-89 min) and was lower than immunocompetent children.

ED clinician survey
Findings from an online survey of clinicians are presented in Table 2.

Discussion
The present study found frequent delay in intravenous antibiotic administration in most children meeting CEC paediatric sepsis criteria that were admitted to hospital with fever.This delay affected a higher proportion of infants compared with preschool-aged children.This frequency of delayed antibiotics was markedly higher than in adults meeting sepsis criteria locally.In the adult population, 48% did not receive antibiotics within 1 h of presentation (CEC report, 2017).However, when we applied SSCM 2020 sepsis bundle recommendations, more than two-thirds of children meeting CEC sepsis criteria received antibiotics within 3 h.This includes all children with serious bacterial infections.This performance is reasonable given that in this cohort, sepsis or septic shock was infrequent.
Early recognition of sepsis in children is challenging for clinicians.Viral aetiology is responsible for a significant proportion of children presenting with systemic illness. 10In the present study, respiratory viruses were isolated in a significant number of children with systemic illness.Clinicians often emphasised the need for time to monitor progress or further tests to elucidate the presence of sepsis.However, the use of laboratory tests supports clinical suspicion of potential sepsis but does not define the diagnosis and was reported to contribute to the delay.The survey response of clinicians suggests that in our environment the balance is towards investigations to clarify the presence of sepsis rather than early treatment to avoid deterioration.This highlights the clinical challenge between early diagnosis and the potential to unnecessarily administer   antibiotics to children without sepsis.Unfortunately, the paediatric sepsis pathway does not assist with clinical reasoning in this instance.This is partly due to the low utility of SIRS criteria in children. 10This has implications on the reliability of the CEC Paediatric Sepsis Pathway.In our retrospective application of the CEC Paediatric Sepsis Pathway, we used a higher threshold of at least four criteria to increase the specificity of our diagnosis.The paediatric sepsis pathway only requires three criteria and therefore more children in this cohort could have met sepsis criteria.The paediatric sepsis pathway needs to be updated to align with new paediatric sepsis definitions to increase reliability and improve recognition and response.
The survey of clinical staff identified ED processes of activity and staff availability as factors impacting the management of sepsis.Several studies cited by the Surviving Sepsis Campaign have reported similar findings. 15,20Factors such as difficulty gaining intravenous access, as identified in the survey, are particular to children, especially unwell infants.Interestingly, febrile immunosuppressed children and those with indwelling central venous access devices received antibiotics more rapidly than others.This is likely due to the availability of intravenous access and clear pathways to treatment in these higher-risk children.
Quality improvement initiatives targeted towards improving ED processes have been shown to be effective in improving paediatric sepsis management. 16,18The reasons leading to a delay or deferral in treating children with signs of systemic illness identified in the present study should be considered when designing interventions aimed at improving performance.

Limitations
Our study is limited by the study design.The small sample size limits the generalisability of our findings.Our sampling strategy introduces bias since we did not include children discharged from ED that may have unidentified early sepsis.The CEC Paediatric Sepsis Pathway criteria, including its reliance on inflammatory markers, are not well validated and limit the interpretation of our findings.Limited documentation viewed retrospectively implies that clinical decisions cannot be clearly elucidated.The online survey partially addresses this.However, there was a low response rate and respondents might not have been directly involved in the clinical management of the cases presented in the present study.

TABLE 1 .
Clinical characteristics of children admitted with fever

TABLE 2 .
Survey of clinicians working in ED