Article Text

Download PDFPDF

Intensive care

Statistics from

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.


L. Jackman, N. Shetty, P. Davies, K. Morris.Birmingham Children’s Hospital, Birmingham, UK

Introduction: We have observed a number of children admitted after surgery with a normal lactate who develop late onset hyperlactatemia (>3 mmol/l) within the following 12 hours. We undertook this study to determine the incidence of and risk factors for late onset hyperlactatemia and whether this group of children have adverse clinical outcomes. We elected to study two cohorts of children, firstly a homogeneous group undergoing a single procedure, the Fontan operation, and secondly a more heterogeneous group of children undergoing a variety of procedures requiring CPB.

Methods: A retrospective study in the paediatric intensive care unit of a tertiary referral centre. Cohort 1: 64 patients undergoing a fenestrated Fontan procedure, mean (SD) age 5.4 (1.9) years. Cohort 2: 84 patients aged 3–16 years (mean 7.8 (3.8) years) undergoing mixed procedures. Operative details were recorded. Arterial blood gases together with lactate and glucose concentrations were recorded at 0, 4, 8, and 12 hours postoperatively. Corresponding hemodynamic variables and levels of inotropic, vasodilator support were collected. Cardiac output was not measured but indirect markers of systemic perfusion were recorded. Information relating to complications, length of mechanical ventilation, and length of PICU and hospital stay were collected. Data from the late onset hyperlactatemia group were compared with the group of children with a normal blood lactate concentration throughout (NOR) using parametric and non-parametric statistics as appropriate (MINITAB).

Results: Cohort 1: 28 patients (44%) developed late onset hyperlactatemia, 14 patients (22%) maintained a low lactate (NOR). A strong association was found between late onset hyperlactatemia and the development of hyperglycemia (p<0.001) and to a lesser extent with the use of epinephrine (p<0.03). Cohort 2: 23 patients (27%) developed late onset hyperlactatemia, 23 patients (27%) maintained a low lactate (NOR). Again a strong association was found between late onset hyperlactatemia and hyperglycemia (p<0.001). Both cohorts: no differences were found with respect to age, length of CPB or circulatory arrest, indirect markers of cardiac output, duration of ventilation, complications, length of PICU, and hospital stay.

Conclusions: Late onset hyperlactatemia is common in children following cardiac surgery requiring CPB. We have found no evidence that it is attributable to tissue hypoperfusion or that it is associated with adverse clinical outcomes. The strong association between late onset hyperlactatemia and hyperglycaemia would support a metabolic process, which could be catecholamine mediated.


P. L. Davies, M. Baird, A. Wills, R. Yates, T. Montague, R. Wynn, K. Hawkins.Royal Manchester Children’s Hospital, Manchester, UK

Aims: Advances in haemopoietic stem cell transplantation (HSCT) have lead to its extended use in children with haematological, metabolic, and immune disorders. Many children become critically unwell following haemopoietic stem cell transplantation and require admission to the paediatric intensive care unit (PICU). Traditionally the outcome of these children has been poor, leading to debate about the provision of intensive care to these children. Some more recent data have indicated an improvement in outcome. The aim of this study was to analyse the outcome of children admitted to PICU following haemopoietic stem cell transplantation in our unit and to identify any factors that would predict outcome.

Methods: The notes of all children admitted to the PICU in our unit as a direct consequence of haemopoietic stem cell transplantation over a 10 year period were retrospectively analysed. Details about the patient’s age, sex, underlying diagnosis, type of transplant, and time from transplantation were recorded. Signs of organ failure were noted, with cardiac failure being defined as the need for inotropic support, respiratory failure as the need for mechanical ventilation, renal failure as a creatinine >100 μmol/l, and hepatic failure as a bilirubin of >40 μmol/l. The worst physiological variables in the first 24 hours of PICU admission were used to calculate APACHE-II scores. Patient survival was documented at PICU discharge and 6 months from discharge.

Results: 32 children were admitted to PICU on 41 occasions. 19 (46%) admissions survived to discharge and 11 (27%) were alive 6 months later. There was no difference in survival based on primary diagnosis with 5/14 (35.7%) with malignant disease and 6/27 (22.2%) with non-malignant disease surviving to 6 months (p 0.46). Type of transplant or time since transplant made no difference to outcome with 7/16 (43.7%) admitted >100 days from transplant surviving compared to 4/25 (16%) admitted <100 days. 3/13 (23.1%) children ventilated over 7 days survived compared with 8/28 (28.6%) ventilated <7 days, which was not significantly different. The number of organs that failed did not significantly alter outcome, this included one child who survived despite all four organ systems failing. The need for inotropic support did not significantly increase mortality (p 0.06), although the need for multiple inotropes was associated with reduced survival (p 0.018). Mean APACHE-II scores were significantly lower in children that survived to 6 months (mean 20.57) compared with those who died (mean 26.45; p 0.032).

Conclusions: Our data reflect the increasing survival of children admitted to PICU following haemopoietic stem cell transplantation. However, identifying which children will survive remains extremely difficult. This suggests that active intensive care support is appropriate in children who become critically ill following haemopoietic stem cell transplantation.


M. G. Davidson, J. C. Coutts, G. Bell.Royal Hospital for Sick Children, Glasgow, UK

Aim: There is an increasing awareness of the role of flexible bronchoscopy in the paediatric intensive care unit (PICU). The aim of the study was to evaluate the role of flexible bronchoscopy in a large PICU.

Method: The first 200 bronchoscopies undertaken in our PICU were reviewed. The results of the bronchoscopies and the microbiological data from broncho-alveolar lavage were analysed.

Results: Two hundred bronchoscopies were undertaken in PICU between August 1990 and June 2003 on a total of 129 patients, which represents 2% of the PICU population during this period. The mean PICU admission rate over the time period was 510 children per annum. The proportion of children in PICU undergoing flexible bronchoscopy has been increasing annually with 29 of the 605 PICU patients (4.8%) in 2002 being bronchoscoped. The mean number of bronchoscopies per child was 1.69. Seventy three (36.5%) bronchoscopies were undertaken while the child was on extra-corporeal membrane oxygenation (ECMO) support. This represents nearly half of all patients receiving ECMO over the study period. Bronchoscopy was undertaken for a variety of diagnostic reasons including possible airway anomaly, persistent atelectasis, and to isolate bacteriae causing ventilator associated pneumoniae. Twenty of the 200 patients (10%) had endotracheal tube mis-placement (endotracheal tube tip at carina or below) despite apparently appropriate placement on chest radiography. Two patients had significant endotracheal tube obstruction. Broncho-alveolar lavage was performed in 117/200 cases (58.5%). A significant positive culture, as deemed by a consultant microbiologist reviewing the data, was found in 34 of the 117 broncho-alveolar lavages (30%) undertaken. Twenty three of 200 (11.5%) bronchoscopies were therapeutic in nature, primarily to clear excess secretions during and following ECMO support.

Conclusion: Flexible bronchoscopy is a safe and effective method of investigating airway problems in children in PICU and of isolating organisms causing ventilator-associated pneumoniae. In the majority of the procedures flexible bronchoscopy contributed to optimal patient care. It is especially useful for both diagnostic and therapeutic reasons in children with complex pathologies who have required cardiac surgery and or ECMO support. We have highlighted the high yield of positive results from flexible bronchoscopy in our PICU. In our experience flexible bronchoscopy is now an integral aspect of patient care in PICU.


B. A. Hameed1, H. E. Miller1, A. L. Curran1, L. Series1, S. A. C. Broere1, C. M. Haworth1, R. J. Mccarter1, P. M. Sharples2.1Kids Head Injury Study Group, Frenchay Hospital, Bristol, UK; 2Bristol Royal Hospital for Children, Bristol, UK

Introduction: Traumatic brain injury is a major cause of admission to hospital and to paediatric intensive care units. The aim of treatment is survival with good quality of life, yet few data exist concerning health related quality of life and its determinants after traumatic brain injury.

Aims: (1) To define health related quality of life 1 year after traumatic brain injury compared with controls and to investigate its determinants.

Methods: Longitudinal prospective study of traumatic brain injury children admitted to hospital compared with non-injured controls matched for age, sex, and socioeconomic status. Health related quality of life was assessed using parent report Pediatric Quality of Life Inventory (PedsQL). Injury severity was assessed by admission Glasgow Coma Score (GCS) (severe, GCS 3–8; moderate, GCS 9–12; mild, GCS 13–15), length of hospital stay (LOS), length of post traumatic amnesia (PTA), and discharge KOSCHI score. Cognitive outcome was assessed by Wechlser Intelligence Scale for Children and psychological outcome by Child Behaviour Checklist (CBCL), Birleson Depression Scale (BDS), Impact of Events Scale (IES), and Revised Children’s Manifest Anxiety Scale (RCMAS). Depressive symptoms in the main carer were assessed by Beck Depression Inventory (BDI).

Results: Fifty traumatic brain injury children and 27 controls were recruited, mean age 10.9 years (SD 4.4) in traumatic brain injury subjects, 10.8 years (SD 4.3) in controls. Traumatic brain injury children had significantly reduced health related quality of life compared with controls, particularly severe/moderate traumatic brain injury (PedsQL total score, p<0.0001). Health related quality of life was reduced in all PedsQL domains (physical p 0.001; emotional p 0.002; social p 0.002; school p 0.001). There was a relationship between PedsQL total score and traumatic brain injury severity (LOS p 0.01), cognitive outcome (verbal IQ p 0.03), psychological outcome (CBCL total problem score p<0.0001, externalising index p<0.0001, internalising index p<0.001, total competence p 0.001, and BDS p 0.005), depressive symptoms in main carer (BDI p 0.003), and PedsQL total score 1 month after traumatic brain injury (p 0.02). There was no correlation between 1 year PedsQL total score and age, sex, PTA, discharge KOSCHI, performance IQ, IES, or RCMAS (p>0.05).

Conclusion: Health related quality of life is reduced 1 year after traumatic brain injury. Health related quality of life is related to injury severity, cognitive outcome, and to psychological outcome in child and main carer. Increased psychosocial support post traumatic brain injury may improve health related quality of life.


P. Ramnarayan1, M. Hayden2, F. Craig1, A. J. Petros1, C. M. Pierce1.1Great Ormond Street Hospital for Children, London, UK; 2Bristol Royal Hospital for Children, Bristol, UK

Introduction: End of life decision making and limitation of care is a vital clinical skill. Evidence suggests that a large proportion of in-hospital deaths now occur on the intensive care unit (ICU). This has implications for staff training as well as for resource availability and allocation within critical care units and paediatric wards. This study examines trends in hospital mortality at a large tertiary referral children’s hospital.

Methods: Data regarding date and place of death for all children who died in hospital between April 1998 and March 2004 were extracted from the hospital database. Hospital and ICU (paediatric, neonatal surgical, and cardiac) admission data were also analysed over the same period. For a subset (Nov 1998 to Oct 1999), information regarding do not attempt resuscitation (DNR) status was collected from hospital records. χ2test was used to test statistically significant differences in proportions.

Results: During the study period, 15 250 of 139 172 hospital admissions were to the combined ICUs (10.9%); year wise analysis suggested a fall in this rate (p<0.01). 845 out of 976 hospital deaths occurred in the ICU environment overall from 1998 to 2004 (86.6%), with a year wise increase in this proportion (71.5% in 1997/98; 90.5% in 2003/04; p<0.05). Although hospital and ICU mortality rates did not vary significantly (p>0.1), the proportion of patients who died on ICU after internal admissions from wards increased over the years (15.5% in 1998/99 v 24.1% in 2003/2004; p>0.1). Only 27% of patients who died on ICU from Nov 1998 to Oct 1999 had a previous DNR order. Care was either limited or withdrawn in 80% of ICU deaths during this period.

Conclusions: A large proportion of hospital deaths now occur in an ICU, rather than on the ward to which children were admitted. Non-ICU staff are increasingly becoming unfamiliar with managing death. The involvement of a combined outreach ICU and palliative care team at an early stage will help identify sick children who may not benefit from intensive care.


E. T. Clarke, K. Chetty, S. Edees.Royal Berkshire Hospital, Reading, UK

Introduction and Aim: National guidelines recommend that all hospitals providing care for children have arrangements for those requiring high dependency care.1 We present data on high dependency requirements and case mix in a district general hospital serving 100 050 children aged ⩽16 years.

Methods: Data were collected prospectively for all children admitted to the high dependency unit between April 02 and March 04. Further information was obtained by analysing the case notes retrospectively. Each case was assessed against national recommendations for high dependency care.2 A sample of case notes for children admitted to the general paediatric wards was also examined to ensure children warranting high dependency care were not being missed.

Results: There were a total of 443 episodes of patients admitted to the high dependency unit over the 2 year period. This represents a demand of 2.2 admissions per 1000 children per year. It equates to 5.8% of the 7630 children admitted to the hospital acutely or electively over the same time period. Infants <1 year accounted for 53% of the total high dependency care time. The mean stay on the high dependency unit was 2.1 days (median 1.0, inter-quartile range 0.6 to 2.1). Expecting 60% occupancy, 2.1 high dependency beds per 100 000 children would be required. However, aiming to meet the demand for beds 95% of the time, assuming a Poisson distribution, we calculate that 4.5 beds are required per 100 000 children. The most common factors defining the need for high dependency care were 1) oxygen requirement greater than 50% or NCPAP for bronchiolitis (14%); 2) prolonged or recurrent convulsions (13%); and 3) asthma on intravenous drugs or hourly nebulisers (8.5%). We also demonstrate seasonal variation in service demand with December accounting for 12.4% of the total workload.

Conclusion: The figures provide a basis for the planning of high dependency services for children. They also demonstrate the case mix, age distribution, and seasonal variation in high dependency care in a district general hospital in the UK.




G. Vemuri, S. D. Playfor.Royal Manchester Children’s Hospital, Manchester, UK

Introduction: Over recent years there have been increasing concerns regarding an increase in the number of futile and inappropriate admissions to paediatric intensive care units (PICUs) in the UK.

Methods: A postal questionnaire was sent to the directors of all UK PICUs identified through the Paediatric Intensive Care Society. Respondents were asked to give details of all patients on their unit at the time including age, reason for admission, and any pre-existing medical conditions. Finally an assessment was made by respondents of whether the care being provided in each case was, in their opinion, either futile or inappropriate. All questionnaires were completed on the same day in order to obtain a snapshot of current practice. The following definitions were used:

  • futile treatment: the care being provided will not have the desired outcome or accomplish its intended goals (no physiological effect)

  • inappropriate treatment: treatment is extremely unlikely to be beneficial, is extremely costly or is of uncertain benefit.

Results: We received responses from 21 units (68%) who reported the details of 111 patients. Care was felt to be futile in nine of these cases (8%) and inappropriate in a further 14 cases (13%). Futile cases were most commonly admitted with respiratory failure and all had pre-existing medical conditions, most commonly developmental delay. Where care was felt to be inappropriate, respiratory failure was the most common reason for admission and 93% had a pre-existing medical condition, most commonly cardiovascular disease.

Conclusion: The care provided in 21% of the PICU cases described in this study was felt to be either futile or inappropriate by the directors of those units. Doctors have an ethical obligation to avoid those treatments where there is no net benefit to the patient. Futile and inappropriate PICU admissions could easily fall into this category.


M. Smith.Queen’s Medical Centre, Nottingham, UK

Background: Doctors training in paediatric intensive care medicine come from backgrounds of either anaesthesia or paediatrics, having markedly different prior experience both of clinical medicine and educational programmes. Anecdotally it is suggested that different medical specialties attract different personality types. This study hypothesised different specialties may also therefore attract trainees with different learning styles and that knowledge of them might help improve the teaching programmes in paediatric intensive care medicine.

Method: Three specific questions addressed were:

  • What are the learning styles of doctors in training in paediatric intensive care medicine?

  • Are there any common features among the learning styles of trainees and any differences between background specialties?

  • Can these theoretical learning styles be exploited to improve the efficiency and efficacy of the training provided?

A learning styles questionnaire was administered to a cohort of anaesthetic and paediatric trainees and the results analysed using descriptive statistics and by profile analysis to look for differences between the two groups. In addition, a number of paediatric intensive care medicine trainees were interviewed about their experiences of training, both positive and negative, and to generate suggestions of preferred practical teaching strategies.

Results: Both groups of trainees tended to prefer a reflective style of learning, followed by pragmatic and theorist styles. Anaesthetists were more likely to have very positive or very negative preferences for any given style than paediatricians. The views expressed in the interviews corresponded well with the results of the questionnaire.

Conclusion: It is concluded that educational strategies based on reflection on, and analysis of, real clinical problems by trainees, are the preferred methods of learning of trainees and are in keeping with their learning styles. To implement these, a change of mindset away from a traditional programme of didactic lectures and “teaching rounds” is required.


S. A. Russ3, S. W. Hancock1, M. Quinton1, R. Moore1, P. Harrison2.1Sheffield Children’s Hospital Retrieval Service, Sheffield, UK; 2Princess Royal Spinal Injuries Unit, Northern General Hospital, Sheffield, UK; 3Department of Accident and Emergency, Rotherham District Hospital, Rotherham, UK

Aims: Failure to immobilise patients at risk of spinal cord injury may have devastating consequences. Previous studies in the UK have looked at numbers of at risk patients immobilised,1 but not methods of immobilisation during transfer (movement of patients between hard surfaces in close proximity) and transport (patient movement between facilities). This study reviews current UK practice.

Methods: Postal questionnaires were sent to the retrieval coordinators in 18 UK PICUs asking about methods of spinal immobilisation during transfer and transport of paediatric trauma patients and existence of guidelines for management of that population.

Results: There was a 100% (18/18) response rate (postal plus telephone follow up for two centres). 27% (5/18) of retrieval services employed practice guidelines. For patient transfer 27% (5/18) of retrieval services utilised a patslide device alone and 50% (9/18) utilised a patslide in combination with a vacuum mattress and/or spinal board. For patient transport 67% (12/18) of services had a consistent approach. Devices used were: vacuum mattress alone (4), spinal board alone (2), vacuum mattress + spinal board (1), spinal board + padding (4), trolley alone (1). A spinal board, either alone or with padding, was used by 67% (12/18) of services for at least some of their patient transports. 100% of retrieval services used the traditional triad of hard collar, blocks, and tapes for cervical spine immobilisation.

Conclusion: There is wide variation in the method of immobilisation utilised by UK retrieval services. This may be partly explained by the lack of best practice guidelines. Despite evidence that spinal boards only have a role during extrication2 they continue to be used frequently. Vacuum mattresses, which may confer advantages in terms of patient safety and comfort,3 are not consistently in use for patient transport in the UK.





G. A. Colville, C. Pierce.Great Ormond Street Hospital for Sick Children, London, UK

Aims: Adult intensive care patients suffer from significant rates of post traumatic stress after admission, with symptoms relating both to distressing memories of the admission,1 and memories of delusional experiences such as nightmares and hallucinations.2 Very little is known about the memories or psychological wellbeing of children after an admission to intensive care.

Method: In this study, 39 children were interviewed 2 months after discharge about their memory for events during admission. They and their parents also completed questionnaires about their current psychological state.

Results: In all, 9/39 (23%) children and 20/36 (56%) parents scored above accepted cut offs, suggesting they were at risk of developing post traumatic stress disorder. The correlation between parent and child scores was 0.40 (p 0.01). No association was found between child’s level of post traumatic stress symptoms and their age, sex, or length of stay. However, those children reporting delusional memories had higher scores than those reporting factual memories only (p 0.03).

Conclusion: In this study both children and parents reported significant levels of post traumatic stress 2 months after admission to paediatric intensive care units. Children reporting delusional experiences, such as those associated with withdrawing from sedative medication, were affected to a greater degree than those who did not.

Funded by The Health Foundation.




F. Craig, C. Pierce, M. Peters, P. Ramnarayan.Great Ormond Street Hospital for Children, London, UK

Aims: Children with conditions that are known to be fatal in childhood continue to be admitted to paediatric intensive care units (PICUs). The aim of this study was to determine if a palliative care service can offer an alternative approach to PICU management that is acceptable or preferable to the child and family.

Method: A retrospective case note review of PICU patients referred to a palliative care service over a 2 year period.

Results: Twelve patients were referred to the palliative care service with a diagnosis of a condition that was known to be fatal in childhood. The diagnosis and prognosis was known prior to PICU admission in 10 of these children. The palliative care service was involved only after the children were admitted to the PICU and all families, at that stage, were offered symptom directed treatment on the PICU, at home, in a children’s hospice, or on the hospital ward, as an alternative to PICU management. Five families chose to continue intensive care management and four of these children died in the PICU while ventilated. Seven families chose symptom directed management. Only one of these families chose to stay in the PICU until the child died, although a second child died just before transfer home. Three children died at home (one on withdrawal of ventilation) and one child died on a hospital ward. Two children are alive a year after referral and continue to receive support from the palliative care service. Only one of these families has chosen to continue intensive management.

Conclusions: Children continue to be admitted to a PICU during the end stages of conditions that are known to be fatal in childhood, yet over half the families admitted chose to discontinue intensive care when offered an alternative. A combined outreach PICU and palliative care team could help identify children who may not benefit from PICU admission and ensure that families are offered alternative management options that they may find preferable.


P. Nayak, A. Bagga, S. Nichani.University Hospital of Leicester Glenfield Hospital, Leicester, UK

Aim: To study the safety of 6% pentastarch and compare it with the colloid currently used (that is, 4.5% albumin) in our paediatric intensive care unit(PICU; quaternary paediatric cardiac ICU) when used for volume replacement after cardiac surgery. There is significant controversy as to which is the “best” colloid to use for volume replacement in critically ill children. 4.5% Albumin is one of the most commonly used colloids but has significant cost implications as well as being a blood product. Pentastarch, which is a complex starch compound, has some proven advantages such as reduced cost as well as theoretical advantages such as a reduction in capillary leak and reduced “sludging” in the microcirculation.

Methods: 100 paediatric postoperative cardiac patients were randomised to receive 4.5% human albumin solution (HAS) or 6% Pentastarch. The effects of these two fluids on the haemodynamic (heart rate, blood pressure, and central venous pressure) and haematological parameters (haemoglobin, platelets, fibrinogen, APTT ratio, thrombin time, INR, urea, creatinine, sodium, potassium, and albumin) were compared and analysed over 72 hours postoperatively. The parameters were measured twice per day over this time period. The results were analysed using Mann-Whitney as a test for non-parametric data.

Results: The results show that there is no significant difference (p>0.05) between the use of 6% Pentastarch or 4.5% HAS in post-operative paediatric cardiac patients for any of the haemodynamic or haematological parameters.

Conclusion: Pentastarch is a safe colloid to use in the post-operative paediatric patient with reference to haemodynamic parameters as well as haematological parameters.


B. N. Ofoegbu, S. D. Playfor.Royal Manchester Children’s Hospital, Manchester, UK

Aims: Physical restraint techniques may be used in the critical care environment to prevent treatment interference and to protect patients in some circumstances; primarily to prevent self-extubation. The use of such techniques is not without risk to psychological and physical wellbeing, and remains controversial. The American College of Critical Care Medicine has recently produced clinical practice guidelines in this area. We have conducted a survey to define the current clinical practice in the UK.

Methods: A postal questionnaire was sent to a named senior nurse on all 32 UK paediatric intensive care units (PICUs) identified through the Paediatric Intensive Care Society. Questions involved the use of physical restraint techniques, specifically manual holding of patients, the use of cross-joint splints, and of tying wrists to bedframes. Consent issues surrounding the use of physical restraint techniques were also addressed.

Results: We received responses from 28 units (88%) of whom 19 (68%) said that physical restraint techniques were used. Ten units (36% of those responding) reported the use of manual holding of patients, 14 units (36% of those responding) reported the use of cross-joint splints while one unit secured wrists to bedframes with “posey” restraints. Verbal consent was obtained by 53% of units prior to the application of physical restraint techniques, no units obtained written consent.

Conclusion: The majority of UK PICUs currently use physical restraint techniques, most commonly using cross-joint splints. There is a clear lack of prospective, high quality evidence of the relative risks and benefits involved in physically restraining critically ill children and how these techniques fit into a comprehensive programme of analgesia and sedation.


T. Y. M. Lo1, P. A. Jones1, I. R. Chambers2, J. Croft2, G. Wilson2, T. F. Beattie3, R. Forsyth2, A. D. Mendelow2, R. A. Minns1.1University of Edinburgh, Edinburgh, UK; 2Newcastle General Hospital, Newcastle-Upon-Tyne, UK; 3Royal Hospital for Sick Children, Edinburgh, UK

Aim: We aim to determine whether critically ill head injured children in possession of the different apolipoprotein E (APO E) alleles experience different quantities of secondary cerebral perfusion pressure (CPP) insults, and whether this relates to outcome.

Methods: Sixty five critically ill head injured children and 160 age and sex matched controls were recruited into a two centre prospective case control study. Buccal smear was collected to extract DNA for APO E genotyping. Head injured children also had acute minute by minute physiological recordings down loaded from the intensive care bedside monitors to identify the frequency and duration of age-specific derangements. The total burden of CPP insult was determined by calculating the age related cumulative pressure time index (CPT). Outcome was assessed at 6 months post injury using the modified Glasgow Outcome Score (GOS). The relationship between APO E genotypes and the age related CPT was analysed using the Kruskal Wallis Test.

Results: The distribution ratio of the APO E alleles was similar between the controls and head injured children but a significantly higher proportion of APOE e2 allele was identified compared with previous adult data (χ2 test; p<0.001). No significant difference was found in outcome among the different allelic groups. Among head injured children who had a good recovery, the median CPT value (for CPP) varied significantly among the three allelic groups with the lowest insult value observed in those with one or more e4 allele (6.6 times less than the expected value for the whole group, p<0.05). The median CPT value for APO E 4 possessors with poor outcome was 2.2 times less than the expected value for the whole group suggesting these children had poor outcome despite having suffered less than the expected amount of CPP insult.

Conclusions: Critically ill head injured children with different APO E alleles had significantly different amounts of cerebral perfusion pressure insult (as measured by the cumulative pressure-time index) with the least amount found in APO E e4 allele possessors.


E. O’Dell, S. Tibby, A. Durward, I. Murdoch, J. Aspel.Guy’s Hospital, London, UK

Aims: Base deficit (BD) quantifies a metabolic acidosis without defining aetiology. Factors contributing to a metabolic acidosis include lactic acid and unmeasured anions (together known as tissue acid), hyperchloraemia, and plasma albumin concentration. An equation exists to quantify the effect of albumin on BD; we have recently derived a similar equation for the effect of chloride. By adjusting the BD for these two factors, we hypothesised that the corrected BD would quantify tissue acid.

Methods: Sixty patients admitted to the intensive care unit with meningococcal sepsis over 2 years were studied, median (interquartile) weight 13 kg (10 to 20), median PIM risk of mortality 10.3% (6 to 16). Arterial blood was drawn at admission, and at 4, 8, 12, 18, 24, and 48 hours. BD was calculated from a standard algorithm in the blood gas analyser, and tissue acid via the Stewart-Fencl method.

Results: A substantial metabolic acidosis was present at admission (mean pH 7.31, SEM 0.01, mean BD −8.9, SEM 0.5), which persisted at 24 hours. The admission BD was predominantly due to tissue acid, however, hyperchloraemia was a significant contributor by 8 hours. This effect was related to the amount of chloride administered in the resuscitation fluids (r2 = 0.41). The BD was weakly associated with total tissue acid (r2 = 0.27), however this improved dramatically after BD correction for chloride and albumin (r2 = 0.83).

Conclusion: It is now possible to partition the BD at the bedside, quantifying the contribution of tissue acid, chloride and albumin to a metabolic acidosis.


L. Edwards, A. Siddiqui, D. Harrington, K. Morris.Birmingham Childrens Hospital, Birmingham, UK

Introduction: The Sano modification of the Norwood procedure incorporates a right ventricle-to-pulmonary artery (RV-PA) conduit as the source of pulmonary blood flow, replacing the modified Blalock-Taussig (BT) shunt of the classical Norwood procedure. It has been suggested that the Sano modification will improve the early post-operative haemodynamic profile, through optimisation of the balance between pulmonary and systemic circulations and, as a result of improved coronary perfusion during diastole, improved myocardial performance.

Aim: To compare early postoperative haemodynamic performance between patients undergoing a Norwood procedure with a modified BT shunt or an RV-PA conduit.

Methods: A retrospective study of 51 infants with hypoplastic left heart syndrome undergoing Norwood surgical palliation with a BT shunt (group 1; n = 23) or an RV-PA conduit (group 2; n = 28) between May 2001 and June 2003. Details of surgical technique and operative variables were recorded. Arterial and venous blood gases were sampled at 10 set time points in the initial 96 hours postoperatively, allowing inferences to be made about relative pulmonary and systemic blood flows. Corresponding haemodynamic variables and levels of inotropic, vasodilator support were collected. Biochemical markers of renal and liver function over five post-operative days were used as additional surrogate markers of systemic oxygen delivery.

Results: Circulatory arrest time was longer in group 1 (46 mins (11–70) v 23 mins (6–65); p 0.003), but cross clamp (49 mins (38–96) v 61 mins (12–76); p 0.001) and cardiopulmonary bypass times (69 mins (45–322) v 106 mins (45–138); p 0.001) were shorter than group 2. Diastolic BP was significantly lower in group 1 (p<0.001) with a trend towards a higher systolic BP but no difference in mean BP. No between group differences were found in PaO2, PaCO2, SaO2, FiO2, or PaO2:FiO2 ratio or in blood lactate, oxygen extraction ratio, or estimated Qp:Qs. Serum creatinine, urine output, need for dialysis and serum ALT, and bilirubin did not differ between groups.

Conclusion: With the exception of a higher diastolic blood pressure in the Sano group, we found no difference in the early haemodynamic profile, markers of pulmonary blood flow, markers of systemic oxygen delivery, or end organ function in patients undergoing an RV-PA conduit rather than a BT shunt.


M. Jayashree, S. Singhi.Advanced Pediatric Centre, Postgraduate Institute of Medical Education and Research, Chandigarh, India

Aim: To determine the incidence of hyponatremia in critically ill children receiving conventional maintenance fluids and to identify the risk factors for the same.

Materials and Methods: This was a prospective observational study conducted in the paediatric intensive care unit (PICU) of a tertiary care, teaching, and referral hospital, from mid March to mid April 2004. During the study period 38 patients with a median PRISM score of 13 (range 4–20) were admitted to the ICU. In all patients serum and urinary sodium (by autoanalyser) and osmolality (by freezing point method) were measured twice daily. Daily total intake of oral and intravenous fluids, net daily intake of sodium (mEq/Kg) and electrolyte free water (EFW, ml/kg), total urine output (ml/kg), urinary sodium excretion (mEq/kg), and net balance of fluid and sodium were calculated. Hyponatremia was defined as serum sodium ⩽130 mEq/l. Patients who had one or more episodes of hyponatremia were identified and compared with the normonatremic group to identify the risk factors using Mann-Whitney U and χ2 tests. Within group comparison was done with paired t test and Wilcoxon ranks test.

Results: Fourteen episodes (3.5 episodes/100patient days) of hyponatremia were recorded in 12 patients over 397 patient days. One patient was hyponatremic at admission, while the rest occurred during ICU stay. The median length of ICU stay prior to hyponatremia was 3.5 days (range 1 to 15 days). The serum sodium in these patients dropped from a median of 139 (range 154 to 124) to 128 (range 127 to 129) mEq/l. The median serum osmolality and urine osmolality were 276 (range 260 to 329) mOsm/Kg and 340 (range 180–649) mOsm/Kg, respectively. The median net balance of sodium and fluid was −5.9 (range −14.4 to +7.60) mEq/Kg and 41 (range 5.6 to 179) ml/kg, respectively. The median EFW intake was 79 (range 24 to 115) ml/kg/day. There was no significant difference between the hyponatremic and normonatremic patients with respect to the median sodium (3.9 v 4.9 mEq/kg) and EFW intake (79.0 v 89.1 ml/kg) and the sodium excretion (7.0 v 8.0 mEq/kg). Similarly in the hyponatremic group, there was no significant difference between the median sodium (3.9 v 2.9 meq/kg), and EFW intake (79.0 v 91.1 ml/kg) and sodium excretion (7.9 v 8.9 mEq/kg) on the day preceding and following hyponatremia. The observed fall in sodium could not be explained by the total fluid given in the hyponatremic group.

Conclusions: In our ICU, incidence of hyponatremia in patients on normal maintenance fluids was low. The hyponatremic children were hypo-osmolar with a negative sodium balance due to natriuresis. However, magnitude of EFW intake, natriuresis, and negative sodium balance in hyponatremic and normonatremic patients was similar. Our findings do not support the contention that use of conventional “hypotonic maintenance fluids” aggravate incidence of hyponatremia.


H. L. Livingstone1, S. Whiteley2, J. B. Luntley3.1Leeds University, Leeds, UK; 2St James University Hospital, Leeds, UK; 3Alberta Childrens Hospital, Calgary, Canada

Aims: Use of the modified early warning score has been shown to improve outcome in critically ill adult patients.12 No widely accepted paediatric equivalent currently exists. We have carried out a Delphi study to determine risk factors for development of critical illness in children and used this data to construct a paediatric advanced warning score.

Methods: Following a review of the literature to identify potential risk factors for the development of critical illness in children we constructed a Delphi questionnaire to determine the perceived weighting of these risk factors together with normal, abnormal, and highly abnormal age specific ranges for a number of physiological parameters. Questionnaires were sent to 56 consultants from various paediatric sub-specialities. The most highly ranked variables were then used to construct a paediatric early warning scoring tool and response matrix.

Results: Three Delphi rounds were required to achieve satisfactory consensus. Eight factors from a patient’s clinical history were identified as potentially significant risk factors. These included neutropenia/immunosupression, pre-existing cardiac, neurological, or haematological oncological disease, prematurity, time since hospital admission, and pre-existing central venous access. Clinical signs and/or physiological variables considered useful by the panel in detecting the onset of critical illness included respiratory and heart rate, capillary refill time, signs of respiratory distress, neurological status, and temperature. From this information, a prototype age group specific, paediatric early warning score was constructed. Colour coded normograms of physiological variable are used to define the risk of potential critical illness and the identifiable risk factors used to weight the response, which may range from a continuation of existing care, to immediate admission to the paediatric intensive care unit.

Conclusion: The score is currently being prospectively validated to determine its sensitivity, specificity, and clinical usefulness.



Abstract G86, figure 1, BD uncorrected and corrected.

Abstract G86, figure 2, BD uncorrected and corrected.