Article Text

Download PDFPDF

Intensive care

Statistics from


S. Harigopal1, K. Thorburn1, N. Taylor2, H. K. F. van Saene2. 1Department of Paediatric Intensive Care; 2Deparment of Microbiology, Royal Liverpool Children’s Hospital–Alder Hey, Liverpool, UK

Introduction: The use of antibiotics in severe bronchiolitis has long been debated. There are no studies to date on the incidence of bacterial co-infection in severe bronchiolitis.

Objectives: To determine the incidence of bacterial co-infection in RSV + bronchiolitis in a tertiary paediatric intensive care unit and to study the impact of co-infection in this patient group.

Methods: Prospective microbiological analysis of endotracheal aspirates on all RSV + bronchiolitis patients on admission to the PICU.

Results: A total of 44 patients, all ventilated, were studied. Median age 2.0 months (IQR: 1.1–5.1). 24 received antibiotics before admission. 17 (39%) cultured bacteria from the endotracheal aspirate on admission to the paediatric intensive care unit: 7 (16%) infected, 10 (23%) colonised. 10/17 (59%) had comorbidities (congenital heart or chronic lung disease, immunodeficiencies, abnormality of large airways) as compared to 11/27 (37%) in RSV only group. There were 2 deaths, 1 in the co-infection group (pertussis, requiring ECMO) and 1 in the RSV only group (oncology patient).

Conclusions: Prior antibiotics may have influenced the colonisation group. The incidence of bacterial co-infection in patients admitted to the PICU with severe RSV bronchiolitis was between 16% and 39%. This supports the use of prophylactic antibiotics in this cohort of severe RSV disease.

Abstract G58 Comparison of patients with bacterial co-infection/colonisation and RSV only (median (IQR))


A. Diaz-Caneja, J. Gledhill, T. Weaver, S. Nadel, E. Garralda. St Mary’s Hospital, London

Aims: To explore experiences of parents and children during admission to hospital either paediatric intensive care unit (PICU) or general paediatric ward (GPW); to increase understanding of nature of stress of admission and aspects, which may be particularly difficult for children and parents.

Methods: Twenty parents whose children had been admitted to hospital. Purposive sampling method to ensure representation of both groups (PICU and GPW). Data collected using a semi-structured interview. Interviews were recorded and transcribed. Content analysis was carried out. A coding frame was developed containing the themes, which appeared relevant to the aims. The coding frame was elaborated and modified as new themes and sub-themes emerged in the course of the analysis.

Results: There were differences in the experiences in both groups. In PICU, parents described illness related factors (rapid onset, intense reaction, child’s appearance) as the most stressful, followed by environmental factors (quiet atmosphere, medical equipment). In the GPW parents were more concerned about environmental factors. Both groups, however, reported as very helpful good communication and being informed by the medical team, and parents being seen as part of the team. If this did not happen, parents found the experience more stressful.

Conclusions: Admission to hospital is stressful for parents, mainly in PICU. The qualitative nature of the study gives the opportunity to express in detail the parents’ opinions and experiences. These could help develop strategies to support families in similar situations.


P. Ramnarayan1, H. Vyas2, S. Wilson2, P. Maheshwari1, J. Britto1 the ISABEL study group 1St Mary’s Hospital, Imperial College, London, UK; 2Queen’s Medical Centre, Nottingham. UK

Introduction: ISABEL, a novel paediatric differential diagnostic tool that searches medical content, was tested for its utility in critical care. In this setting, important diagnoses may already have been excluded by multiple clinicians, although the impact of providing additional significant diagnostic reminders may be greater.

Methods: A designated fellow from each of five paediatric intensive care units (PICU), two UK and three US, collected data from medical patients with no clear diagnosis at admission (age, clinical features, team differential diagnosis at first contact, and discharge diagnosis; study period May to July 2003). Presenting clinical features were entered into ISABEL after patient discharge. From the ISABEL diagnostic reminder list, a designated consultant from each PICU highlighted additional significant diagnoses not already included in the team differential. Proportion of cases in which ISABEL presented at least one additional significant diagnosis (utility), >50% of the team differential diagnosis (safety), and discharge diagnosis (accuracy) were measured.

Results: 206 children were enrolled (69 UK, 137 US). ISABEL presented at least one additional “significant” diagnosis in 41.3% (85/206 cases), displayed >50% of the team’s initial differential diagnosis in 56.8% (117/206) cases, and the final diagnosis, based on initial clinical features, in 53.4% (71/133 cases).

Conclusions: By reminding critical care teams of useful alternate diagnoses, the ISABEL differential diagnostic tool has the potential to serve as an adjunct to standard diagnostic decision making, especially in PICUs working with smaller or junior teams.


H. Betts1, R. Sherwood2, A. Duffy2, S. Nadel1. 1Department of Paediatrics, Imperial College Faculty of Medicine; 2Department of Chemical Pathology, King’s College Hospital

Objectives: Meningococcal disease (MD) may be associated with raised intracranial pressure, intracranial haemorrhage (ICH), and poor neurological outcome. S100B protein has been postulated as a reliable blood marker of intracranial injury prior to development of clinical signs, levels of >0.6 mg/ml being associated with neurological injury. We aimed to assess whether serum S100B in children with MD correlated with neurological events and outcome.

Method: Retrospective analysis of serum S100B level in children with MD and correlation with outcome. Serum S100B level was measured by ELISA in 65 children (age 2 months to 17 years) with MD on admission to PICU and in convalescence. Children were grouped by severity of illness using PRISM II.

Results: Acute serum S100B ranged between 0.08 and 15.6 μg/ml (median 0.31). In children with predicted mortality (PM) <10%, median serum S100B was 0.24 μg/ml. In those with PM 10–50%, median serum S100B was 0.37 μg/ml; and in those with PM>50%, median serum S100B was 0.88 μg/ml, (p<0.05). There was no correlation with level of serum S100B and ICH or raised intracranial pressure during PICU admission. In children with meningococcal meningitis alone, serum S100B did not predict neurological injury (16/19 children with serum S100B in the normal range (<0.2 μg/ml) developed acute signs of intracranial injury). There was no correlation between serum S100B and age, peripheral blood white cell count or serum creatinine, factors which may influence serum levels. Two children developed neurological sequelae and three died, two from multiorgan failure and one from brain stem death. Serum S100B level on admission predicted severity of illness, but not neurological outcome.

Conclusion: Serum S100B level correlated with severity of MD but not with neurological outcome, even in children with meningitis alone. Serum S100B may be a useful marker of severity and organ failure but was not useful to identify those at risk of ICH or poor neurological outcome in MD.


T. Y. M. Lo1,2, P. A. Jones1,2, I. R. Chambers3, A. Clark3, G. Wilson3, J. Croft3, B. Fulton3, R. Forsyth3, P. J. D. Andrews1,4, A. D. Mendelow3, R. A. Minns1,2. 1University of Edinburgh; 2Royal Hospital for Sick Children, Edinburgh; 3Newcastle General Hospital, Newcastle-upon-Tyne; 4Western General Hospital, Edinburgh

Aims: We aimed to identify the best predictor of outcome in paediatric traumatic brain injury by (1) quantifying the physiological derangements using pre-set age specific theoretical thresholds, (2) relating these values and a cumulative pressure time index to outcome, and (3) determining age related critical cerebral perfusion pressure (CPP) thresholds.

Methods: Seventy nine head injured children (age 2–16 years) from two regional centres (Edinburgh and Newcastle-upon-Tyne) with minute by minute recordings of physiological variables within 24 h of injury were studied. The prospective time series data included intracranial pressure, arterial blood pressure, CPP, oxygen saturation, temperature, and heart rate and were analysed to identify abnormal (outside pre-set normal age specific physiological limits) epochs lasting more than 5 min. Outcome was assessed at 6 months post injury and quantified as independent (good recovery, moderate disability) or poor (severe disability, vegetative, or dead). Univariate and multivariate logistic regression modelling was used to determine predictors of outcome and ROC curves employed to find the severity of derangement. The severity of secondary insults of CPP was determined by calculating a cumulative pressure time index (CPT) as the area below predefined age related threshold limits and the CPP curve:


These thresholds were then altered by 10% and 20% to determine whether they might differentiate outcome.

Results: In the final model, only low CPP significantly predicted outcome (p 0.004, OR 0.94, CI 0.91 to 0.98). Those with a poor outcome had a CPT value more than 10 times greater than those with an independent outcome with a specificity of 90% and a sensitivity of 75%. ROC curves demonstrated that the initial thresholds were the best predictor of outcome.

Conclusion: Of the routine monitored variables, CPP was the best predictor of outcome. The severity of CPP derangements as determined by an age related index using magnitude and duration is useful in quantifying secondary insults after TBI in children. Derived CPP critical thresholds for children aged between 2 and 6 years, 7 and 10 years, and 11 and 16 years were reported.


K. Vijayakumar1, R. Knight1, L. Hunt2, N. Kane1,2, P. Murphy2, M. Hayden2, J. Patel1,2, P. M. Sharples1,2. 1Frenchay Hospital, Bristol; 2Bristol Royal Hospital for Children

Background: NTC is an important cause of childhood mortality and morbidity. EEGs are often performed in children with NTC, but few data exist concerning the relationship between EEG findings, neurological state, and outcome.

Aim: (1) To examine the results of EEGs obtained in all children with NTC admitted to a regional PICU over the 5 years 1997–2001. (2) To relate the results to depth of coma, clinical seizures, raised ICP, and outcome.

Methods: A regional PICU database, incorporating clinical, radiological, and outcome data, of all admissions for NTC was interrogated and cross referenced with neurophysiological records. Coma was defined as Glasgow Coma Score (GCS) <13 for >6 h. The background activity of the initial and worse EEGs was classified as: (1) normal, (2) slow waves, (3) generalised low amplitude, (4) burst suppression, or (5) isoelectric. Seizure discharges on EEG were classed as: (1) isolated, (2) electrical storms, or (3) continuous. EEGs were excluded from analysis if the patient was receiving thiopentone.

Results: 275 EEGs were recorded in 130 patients over 141 PICU admissions. 141/330 (39%) total NTC admissions underwent EEG. Median age was 2 years, range 0.0l–16; median GCS was 8. Clinical seizures were seen in 117/130 (90%) patients. Initial EEG background appearances were: normal (25:18%); slow waves (79:56%); burst suppression (13:9%); generalised low amplitude (5:4%); isoelectric (7:5%); and continuous seizure discharges (10:7%). 53/141 (38%) initial EEGs showed seizures. There was a significant relationship between initial EEG background and outcome (p 0.009) and EEG seizures and outcome (p 0.009). Other correlations were EEG background and raised ICP (p 0.01) and EEG seizures and clinical seizures (p 0.02). There was no correlation between GCS and either EEG background (p 0.49) or EEG seizures (p 0.22). Multivariant analysis showed that raised ICP and EEG appearance were independent predictors of poor outcome (p<0.001; p 0.04).

Conclusion: Our results indicate that the EEG has a role in NTC not only in the detection and monitoring of seizures but also as a prognostic indicator.

View Abstract

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.