Search tips
Search criteria 


Logo of archdischArchives of Disease in ChildhoodVisit this articleSubmit a manuscriptReceive email alertsContact usBMJ
Arch Dis Child. 2007 April; 92(Suppl1): A23–A25.
PMCID: PMC2066113


Intensive care

G/TUES/INC1 Heart rate variability in children post cardiac bypass surgery

L. McGlone, N. Patel, M. Danton. Royal Hospital for Sick Children, Glasgow, UK

ObjectiveChanges in heart rate variability (HRV) may be useful predictors of outcome in cardiac disease and following surgery. This is the first study to investigate changes in HRV in children in the immediate 24 h following cardiac surgery.

MethodsTwenty seven children (age one month to 15 years) undergoing cardiac surgery were recruited over a 10‐month period. Complete data were obtained for 20 children, all of whom underwent surgery requiring a period of cardiac bypass. HRV was measured for one hour preoperatively and continuously for 24 h postoperatively. Time domain and frequency domain analyses of HRV were performed.

ResultsTime domain analysis revealed a significant reduction in HRV measures at all time points postoperatively, compared to preoperative values. There was a significant association between one of the time domain measures (SDNNi) and duration of cardiac bypass (r = −0.486, p = 0.035). There was a trend towards a reduction in SDNNi with increasing cross clamp time (r = −0.392, p = 0.097), and reduction in SDNNi with longer PICU stay (r = −0.398, p = 0.092). Frequency domain analysis of HRV demonstrated a significant reduction postoperatively at all time points, in total power and the low and high frequency ranges. This indicates a reduction in both sympathetic and parasympathetic control of the heart. There was a correlation between postoperative total power and cardiac bypass time which approached statistical significance (r = −0.439, p = 0.053).

ConclusionsHRV can be easily measured in the immediate period post cardiac surgery. HRV is significantly reduced post cardiac surgery and reflects a reduction in autonomic cardiac control, which may contribute to low cardiac output states at this time. The mechanisms of this global reduction in autonomic control remain to be elucidated, but are related to length of bypass time. HRV may represent an objective clinical predictor of outcome post cardiac surgery in children.

G/TUES/INC2 Changes in surfactant in children with acute hypoxic respiratory failure

T. D. R. Burrows, C. T. Ryder, N. Salam, F. Groom, C. E. Jones, C. Hawkins, M. J. Marsh, A. D. Postle. University of Southampton, Southampton, UK

AimsSurfactant composition and function are altered in acute lung injury (ALI) and acute respiratory distress syndrome (ARDS). Acute hypoxic respiratory failure (AHRF) represents a milder form of respiratory failure with comparable hypoxaemia without the bilateral chest x ray infiltrate seen in ALI and ARDS. We asked if there were similar surfactant abnormalities in AHRF as have been shown in ALI and ARDS.

MethodBronchoalveolar lavage (BAL) samples were taken on days 1–4, 7 and before extubation from children ventilated for AHRF (PaO2/FiO2 ratio <40KPa). Phosphatidylcholine (PC), phosphatidylglycerol (PG), and phosphatidylinositol (PI) totals and percentages of total phospholipid (PL) were measured by mass spectrometry. PC comprises different molecular species with variations in fatty acids; we compared concentrations of PC species characteristic of pulmonary surfactant, plasma and neutrophil cell membranes. Results were compared with 8 controls. Exclusion criteria included pre‐existing lung disease and chronic ventilation. The study had local research ethics committee approval. All results are expressed as medians. Mann–Whitney U test was used to compare data.

ResultsThirty four children were recruited. Total surfactant PL on day 1 was decreased in AHRF compared to controls (52.1 v 283.0 nmol/ml, p = 0.002), with no significant alteration to fractional concentrations (% total PL) of PC (91.7 v 91.5%, p = 0.8), PG (5.63 v 6.96%, p = 0.6) or PI (2.67 v 1.55%, p = 0.2). All these changes persisted until extubation. Concentrations of PC species characteristic of surfactant were decreased: PC16:0/16:0 (31.3 v 45.3%, p<0.001); PC16:0/14:0 (5.23 v 6.72%, p = 0.018). Those species enriched in plasma: PC16:0/18:2 (9.31 v 6.76%, p<0.001) and neutrophil membranes: PC16:0/18:1 (17.49 v 13.44%, p = 0.002); PC18:0/18:1 (1.97 v 1.27%, p = 0.001) were increased. Importantly, PC species concentrations were not restored to control values by extubation.

ConclusionChildren with AHRF show surfactant changes similar to those seen in ALI, with decrease in the major species of PL. The proportions of PC, PG and PI do not change. The decrease in total surfactant may represent decreased production or increased hydrolysis. Clinical improvement precedes return of surfactant to normal levels.

G/TUES/INC3 Data quality issues and audit of paediatric intensive care using Paediatric Index of Mortality 2

A. Plunkett1, D. Ridout2, G. Oldham1, A. Hoskote1, K. Brown1. 1Great Ormond Street Hospital NHS Trust, London, UK; 2Institute of Child Health, London, UK

AimTo investigate quality differences in data collected contemporaneously by two groups: (a) non‐clinical clerical staff (data submitted to PICANET for national audit), and (b) intensive care consultants (for local database).

MethodsThe project was registered with Research and Development Office.

SettingUK paediatric cardiac intensive care unit with congenital heart surgery, extra‐corporeal support and transplant services.

Data1072 consecutive admissions (June 04–December 05). Parameters compared were: (a) items used to calculate the Paediatric Index of Mortality 2 (PIM2) score, (b) primary diagnosis. From a smaller random sample (5% of total), all parameters were checked against the medical records (assumed to be a “gold standard”).

Results(1) The non‐clinician dataset (NCD) included 100% of admissions. The consultant dataset (CD) included 82%. The missing 18% from CD had significantly lower PIM2 scores than the other 82% (mean 1.2% v 4.7%) p<0.001 and included some high dependency patients under the cardiology team. (2) Where PIM2 scores were available from both datasets (n = 879), mean CD PIM2 score was twice as high as NCD (8.0% v 4.7%) p<0.001. Bland‐Altman 95% limits of agreement for CD: NCD ratio is (0.42, 9.34). (3) In‐hospital mortality was 4.55%. Adjusted mortality ratios (SMRs) based on mean PIM2 scores from the 2 datasets differed: 0.97 (NCD) v 0.57 (CD). (4) Analysing 5% of cases in more detail; CD mean PIM2 score differed significantly from medical records (mean PIM2 for medical records  = 5.8%) (table). (5) Primary diagnosis correctly recorded more frequently in CD (78% v 69%).

Table thumbnail
Abstract G/TUES/INC3

ConclusionsData quality may be influenced by the identity of the person who collects it. At our centre, a clinical background increased the likelihood of identifying the correct diagnosis but seems to have biased PIM2 score toward higher values. The mean PIM2 score may be affected by the presence of highly influential outlying values. The differences in SMRs generated by the two datasets could reflect the impact of such data errors. Regular audit and quality control of national audit data is recommended.

G/TUES/INC4 Hyperglycaemia and outcome in critically ill children with meningococcal sepsis

K. Day, N. Haub, D. Inwald. St Mary's Hospital NHS Trust, London, UK

AimsTo investigate the relation between hyperglycaemia and outcome in children ventilated for meningococcal septicaemia.

MethodsRetrospective case notes review, looking at consecutive children ventilated for meningococcal sepsis and admitted to our PICU over a 3‐year period (2001–4). 97 cases were identified, with a median age of 2.1 years (IQR 1.2–4.7) and a median length of stay of 4 days (IQR 4–7). 4 (4%) died. Data collected included platelet‐neutrophil (PN) product, predicted mortality (based on PRISM score), nosocomial infection, ventilator free days at 30 days and use of CVVH. Mean blood (glucose) was determined for the following three epochs: (1) first 24 h (2) first 48 h and (3) the entire PICU admission. Data were non‐parametric. The Mann–Whitney U (MWU) test was used to compare independent samples and Spearman's correlation was used for correlations.

ResultsOnly 8 patients maintained [glucose] <6.7 mmol/l for the entire duration of their PICU admission. 14 patients received insulin based on [glucose] >10 mmol/l and/or glycosuria. 10 patients required CVVH and 27 patients developed nosocomial infection. Mean [glucose] in all epochs was not significantly different in the patients who developed nosocomial infection (p>0.05, MWU) and in those who required CVVH (p>0.05, MWU). There was also no correlation between mean [glucose] in any epoch and ventilator free days at 30 days. However, as expected, the PN product was significantly lower in patients who developed a nosocomial infection (p = 0.005, MWU) and in those requiring CVVH (p = 0.021, MWU). Predicted mortality was also significantly higher in patients who developed a nosocomial infection (p<0.001, MWU) and in those requiring CVVH (p<0.001). PN product correlated with ventilator‐free days and predicted mortality correlated inversely with ventilator free days (p<0.001 and p<0.001 respectively, Spearman's rho).

ConclusionsHyperglycaemia was not associated with a poor outcome in this group of critically ill children.

G/TUES/INC5 Clinical audit of pre‐PICU management of convulsive status epilepticus: room for improvement

V. Kudumula, N. Kasi, D. Thomas, W. P. Whitehouse. Queen's Medical Centre, Nottingham, UK

AimsTo evaluate the clinical features and emergency pre‐PICU management of children with convulsive status epilepticus admitted to PICU, to identify the factors contributing to the severity and the need for intensive care.

MethodsRetrospective study of children admitted to PICU over 2 years with a diagnosis of convulsive seizures and status epilepticus. We compared the pre‐PICU management with standard APLS guidelines for status epilepticus.

Results36/739 (5%) PICU admissions during the study period were admitted with seizures. 26/32 children beyond the neonatal period, had notes available for audit. Their median age was 14.5 months (range 2.5 months to 13 years). Prolonged febrile convulsion (34%) was the most common aetiology. 6/19 cases transferred by ambulance received prehospital treatment. Only 1/6 received an appropriate dose of diazepam. None of these patients received buccal midazolam. The mean time from seizure onset to time of first emergency antiepileptic drug (AED) dose in those without prehospital treatment was 40 (range 10–90) minutes; significantly delayed. APLS guidelines for status epilepticus was strictly followed in 38%. 96% of the cases received benzodiazepines as the first line medication in hospital. 24% received more than 2 doses of benzodiazepines in total. 25% of the benzodiazepine doses given were inappropriately low. Buccal midazolam was given to one patient in a lower than recommended dose. The sequence of AEDs suggested by APLS guidelines was not followed in 35%. 81% were intubated and ventilated.

ConclusionAs the outcome may be adversely affected by delay in the initiation of emergency AEDs, we recommend increasing the awareness of paramedical staff about the importance of starting early treatment. The use of buccal midazolam, already popular with carers and hospital emergency departments, by paramedics and ambulance crews might facilitate implementation of prompt prehospital treatment. We also recommend continuous education and training of hospital staff about the importance of following APLS guidelines and use of appropriate doses of emergency AEDs.

G/TUES/INC6 Incidence of early and late seizures in children following traumatic brain injury and factors predictive of these

B. Hameed1, H. Miller1, A. Curran1, R. McCarter1, P. Sharples2. 1Kids Head Injury Study, Frenchay Hospital, Bristol, Avon, UK; 2Bristol Royal Hospital for Children, Bristol, Avon, UK

IntroductionTraumatic brain injury (TBI) is a recognised risk factor for acute seizures and later epilepsy. However, surprisingly few data exist in children concerning the incidence of early and late seizures and the risk factors predictive of these.

Aims(1) To define the incidence of early and late seizures in a cohort of children with severe, moderate and mild TBI. (2) To investigate which factors are predictive of early and late post‐traumatic seizures.

MethodsProspective cohort study of TBI children admitted to a Regional Neurosurgical Centre (RNSC). Early seizures were defined as those occurring [less-than-or-eq, slant]48 h after injury; late seizures as those occurring >7 days post TBI. Injury severity was assessed by admission Glasgow Coma Score (GCS) (severe, GCS 3–8; moderate, GCS 9–12; mild, GCS 13–15); duration of post‐traumatic amnesia (PTA); length of PICU stay (PICU), total length of hospital stay (LOS); presence or absence of skull fracture, and penetrating versus non‐penetrating injury. CT brain scans were classified using the Lobato and Traumatic Coma Data Bank systems. Functional outcome at hospital discharge was assessed by Kings Outcome Score for Childhood Head Injury (KOSCHI).

ResultsNinety seven children (42 severe/moderate, 55 mild) were studied, mean age 9.7 years (SD 5.2). Mean follow‐up was 4.5 years, range 1.4–7.2. None had epilepsy prior to TBI. 44/97 (45%) received seizure prophylaxis for [gt-or-equal, slanted]1 week; 42 had phenytoin, 1 phenobarbitone, 1 carbamazepine. 13 (13%) had early seizures, 7 (7%) late seizures. There was a significant correlation between early seizures and sex (p = 0.02); GCS (p = 0.025); mechanism of injury (p = 0.003); and administration of seizure prophylaxis (p = 0.001) but no relation between early seizures and any other variable. There was a significant correlation between late seizures and duration of PTA (p = 0.01); skull fracture (p = 0.006); penetrating injury (p = 0.001); CT scan Lobato classification (p<0.001); and seizure prophlaxis (p = 0.014). There was no relation between late seizures and any other variable, including early seizure occurrence (p>0.05). All 20 children who developed late post‐traumatic seizures had received prophylaxis.

ConclusionThe incidence of post‐traumatic seizures is higher in children than in adults. It is possible to identify children acutely that are at high risk of late seizures. We suggest consideration should be paid in these patients to trialling the early use of new agents, aimed at preventing epilepsy developing.

G/TUES/INC7 Normal ranges for aortic blood velocity measured by transoesophageal doppler monitoring in healthy anaesthetised children

S. Robinson, D. Roland, K. Girling, D. Cara, H. Vyas. Nottingham University Hospitals, Nottingham, UK

AimsThere are very few normal transoesophageal doppler (TOD) data available in sedated children that may be applicable to paediatric intensive care practice. Therefore we wished to establish reference intervals for aortic peak velocity (PV) in healthy children during minimally cardiodepressant anaesthesia. Peak velocity can be used to calculate minute distance (the distance travelled by a column of blood in one minute) which can be used as a surrogate for cardiac output.

MethodChildren aged 2–13 years with no pre‐existing oesophageal or cardiac disease who were scheduled for elective middle ear surgery were recruited. A re‐usable 4 mm diameter oesophageal Doppler probe operating at 4 MHz obtained recordings of aortic (PV) and corrected flow time (FTc) for 5 minutes before surgery. Spearman's rank correlation was used for statistical analysis.

ResultsForty four children were recruited over a 6‐month period with no adverse events or outcomes reported. PV median values were 94 m/s (BSA<0.75 m2, n = 14) v 111.5 m/s (BSA>0.75 m, n = 29), FTc median values were 0.359 sec (BSA<0.75, n = 14) v 0.345 sec (BSA>0.75, n = 29). PV in the descending aorta increased with body surface area (BSA) (ρ = 0.424, p<0.01), with FTc showing a mild decrease with increasing BSA (ρ = −0.334, p<0.05) and no significant change with age. Patient height correlated closest with probe length (ρ = 0.868, p<0.001).

ConclusionsIt is often difficult to make a clinical assessment of cardiac output in children and here we present normal values for well children to be used in the evaluation of TOD data. This may be used to aid cardiovascular decision‐making on the PICU. Previous work has suggested peak velocity is constant in children and falls with increasing age after adolescence. Our data indicate an increasing PV until early adult ranges are reached around 6 years of age or a BSA of 0.75 m2. This may be due to greater arterial elasticity in the younger age group which indicates further study will be useful to look into the effects of increasing BSA on cardiac output.

G/TUES/INC8 Antimicrobial resistance in neurologically impaired children requiring intensive care

M. Jardine, I. Galvez, N. Taylor, K. Thorburn, N. Reilly, R. E. Sarginson, H. K. F. van Saene. Royal Liverpool Children's Hospital Alder Hey, Liverpool, UK

AimsIndividuals with underlying disease such as cerebral palsy (CP) may routinely carry aerobic gram‐negative bacilli (AGNB) in low concentrations. This abnormal carriage will lead to overgrowth if the patient is critically ill and admitted to the intensive care unit. There are no available data on the flora found in neurologically impaired children.

MethodsChildren were included with an established diagnosis of cerebral palsy ventilated for four or more days in our PICU between March 1999 and February 2005. Surveillance cultures of throat and rectum were taken on admission to PICU then twice weekly thereafter. Once identified our unit has a policy of administering selective digestive decontamination (SDD) in order to eradicate carriage of abnormal flora, Diagnostic samples were taken on clinical indication. The four endpoints of this six‐year prospective observational single centre study were carriage of abnormal flora, carriage of antibiotic resistant bacteria, infection rates and mortality.

ResultsFifty three children with a total of 77 admissions were included. 89% of CP children (47/53) carried abnormal flora, among which pseudomonas and klebsiella were the predominant potential pathogens. Amongst the carriers 47% (22/47) had resistant bacteria. During their admission, 35 CP children (66%) developed 86 infections. Lower airways and blood were the two most commonly infected sites, while pseudomonas aeruginosa and coagulase‐negative staphylococci were the predominant infecting organisms. 65% (56/86) of infections were primary endogenous, 21% (18/86) were exogenous and 9% (8/86) were secondary endogenous. Median PICU day of infection was 5 (IQR 2–17.5). The overall infected patient rate was 66%, higher than the background rate of children ventilated on PICU for four more days (42%). 17% of children (9/53) died, compared to our overall mortality of 9.8%.

ConclusionInfection rates and mortality were substantially higher in children with cerebral palsy on our intensive care unit. They depict a major source of resistant bacteria both in the community and in the hospital. This finding is new and impacts on their clinical management. Neurologically impaired patients represent a subgroup of PICU patients that should be isolated. Surveillance cultures are mandatory in all patients to tailor antibiotic therapy minimising infection and antibiotic resistance.

Articles from Archives of Disease in Childhood are provided here courtesy of BMJ Group