Inflammation and psychopathology in children following PICU admission: an exploratory study.

BACKGROUND: Survivors of critical illness in childhood commonly display subsequent psychiatric symptoms including emotional and behavioural difficulties, and manifestations of post-traumatic stress disorder (PTSD). Anomalies in inflammatory profiles are an established finding in these childhood psychiatric conditions. OBJECTIVE: This exploratory study aimed to investigate whether abnormal peripheral blood inflammatory markers measured during paediatric intensive care unit (PICU) admission were associated with psychiatric symptoms after discharge. METHODS: We performed a prospective observational cohort study on 71 children with septic illness, meningoencephalitis and other critical disorders admitted to two PICUs between 2007 and 2010. 3-6 months following discharge, subjects were assessed for global psychiatric risk (ie, presence of emotional and behavioural difficulties on the parental Strengths and Difficulties Questionnaire (SDQ)), and for PTSD risk using the child-rated Impact of Events Scale (IES-8). Inflammatory and related biological markers were transcribed from PICU admission notes (white cell count, lymphocytes, neutrophils, C reactive protein (CRP), platelets, fibrinogen and lactate). FINDINGS: Global psychiatric risk at follow-up was associated with abnormal lymphocyte count during admission (χ2=6.757, p=0.014, n=48). In children with sepsis, partial correlation analyses controlling for age and gender highlighted associations between (i) SDQ scores and low lymphocyte count (r=-0.712; p=0.009, n=14), and (ii) IES-8 score and high CRP levels (r=0.823; p=0.006, n=11). These associations remained after correction for multiple comparisons. CONCLUSION: These results support the hypothesis that acute inflammation may play a role in determining the development of psychopathology following PICU admission. CLINICAL IMPLICATIONS: If the findings are replicated, they may help to better highlight which children are at risk of post-PICU psychopathology and appropriately target follow-up.

Post-traumatic stress after PICU and corticosteroid use.

AIM: To examine the association between corticosteroid use in paediatric intensive care units (PICU) and subsequent symptoms of post-traumatic stress disorder (PTSD). METHODS: The subjects were children aged 8-16 years admitted to PICU with sepsis, meningoencephalitis (ME) and other disorders. Illness information was extracted from case notes; 3-6 months post discharge children completed a PTSD symptom questionnaire (eight-item Impact of Events Scale (IES-8)) assessing intrusion and avoidance symptoms. Saliva samples were also collected for cortisol profile analysis. RESULTS: 53 children completed the IES-8 questionnaires. 33 provided saliva samples. 19 (36%) received corticosteroids. In children with sepsis (n=15), corticosteroid use was associated with significantly lower PTSD intrusion symptom scores. There was a trend towards an association between corticosteroid use and lower evening cortisol levels. There was a comparable but weaker trend in children with ME. DISCUSSION: Corticosteroid use may be associated with fewer PTSD symptoms and lower evening cortisol levels following PICU admission in children with sepsis.

Stress hormones and posttraumatic stress symptoms following paediatric critical illness: an exploratory study.

In this exploratory case-control study, we investigated basal cortisol regulation in 5-16-year-old children, 3-6 months following PICU (paediatric intensive care) admission. This was nested within a study of child psychological and cognitive function; 47 children were assessed alongside 56 healthy controls. Saliva samples were collected three times per day (immediately after waking, waking +30 min, and waking +12 h) over two consecutive weekdays. In addition, data on posttraumatic stress symptoms were ascertained from 33 PICU admitted children using the Impact of Events Scale-8 (IES-8). Primary analysis revealed no significant differences in basal cortisol concentrations between PICU discharged children and healthy controls (p > 0.05). Secondary analysis in the PICU group identified a significant positive association between posttraumatic stress symptoms and evening (waking +12 h) cortisol concentrations (p = 0.004). However, when subject to multivariate analysis, evening cortisol was a modest independent predictor of IES-8 scores, relative to the presence of septic illness and poor pre-morbid health. We conclude that paediatric critical illness does not appear to result in marked perturbations to basal cortisol at 3-6 month following discharge. There was evidence of a link between evening cortisol and symptoms of PTSD, but this was not a robust effect and requires further elucidation.

A supported psychoeducational intervention to improve family mental health following discharge from paediatric intensive care: feasibility and pilot randomised controlled trial.

OBJECTIVE: To assess feasibility and pilot a supported psychoeducational tool to improve parent and child mental health following discharge from a paediatric intensive care unit (PICU), in preparation for a large randomised controlled trial (RCT). DESIGN: Feasibility assessment and single-centre, parallel group, pilot RCT. A concealed computer generated list was used to randomise participants, with an allocation of 2:1 in favour of the intervention. SETTING: A PICU in an acute care hospital in London, UK. PARTICIPANTS: 31 parents of children aged 4-16 years-old admitted to PICU. INTERVENTION: Parents received a psychoeducational tool supported by a telephone call. The psychoeducational tool outlined the possible psychological reactions in children and parents alongside management advice. The telephone call addressed each family's postdischarge experience, reinforced the psychoeducational material and encouraged parents to put into practice the advice given. MAIN OUTCOME MEASURES: The primary outcome was the number of feasibility criteria successfully met (linked to the intervention and the study design). Secondary outcomes were questionnaire data collected at 3-6-month follow-up assessing mental health in parents and children. RESULTS: 31 parents were randomised (intervention n=22; treatment as usual, TAU n=9). 23 parents were included in the analysis of secondary outcomes (intervention n=17; TAU n=6). 3 (of 6) intervention and 1 (of 6) study design feasibility criteria were fully met. All unmet criteria could be addressed with minor or significant modifications to the protocol. At follow-up there was a tendency for parents who received the intervention to report lower post-traumatic stress symptoms in themselves and fewer emotional and behavioural difficulties in their children than TAU parents. This needs to be explored in a fully powered trial. CONCLUSIONS: This feasibility and pilot RCT provided valuable information on the intervention and trial design for a full RCT. TRIALS REGISTRATION NUMBER: NCT01737021; Results.

Persistence of Neuropsychological Deficits Following Pediatric Critical Illness.

OBJECTIVE: To study 12-month persistence of neuropsychological deficits in PICU survivors. DESIGN: Prospective follow-up study. SETTING: Two PICUs. PARTICIPANTS: Children 5-16 years old with neuropsychological deficits 3-6 months following PICU care for meningoencephalitis, sepsis, and other critical illnesses (excluding other primary neurological disorders). INTERVENTIONS: None. MEASUREMENTS AND MAIN RESULTS: Neuropsychological function was assessed using the Cambridge Neuropsychological Test Automated Battery, the Children's Memory Scale, and the Wechsler Abbreviated Scale of Intelligence or Wide Range Intelligence Test. Forty-seven of 88 PICU admitted children (53%) were identified as neuropsychologically impaired 3-6 months after discharge; of these, 23 provided 12-month follow-up data. In spite of significant improvements in measures of memory, there was little change in intelligence quotient and visual attention over the study period, and children's educational progress remained below expectation. CONCLUSIONS: We found persistently reduced neuropsychological function following PICU admission in the critical illnesses under study.

Mental and physical well-being following admission to pediatric intensive care.

OBJECTIVE: To assess mental and physical well-being in school-aged children following admission to pediatric intensive care and to examine risk factors for worse outcome. DESIGN: A prospective cohort study. SETTING: Two PICUs. SUBJECTS: A consecutive sample of 88 patients 5-16 years old (median age, 10.00 yr; interquartile range, 6.00-13.00 yr) admitted to PICU from 2007 to 2010 with septic illness, meningoencephalitis, or other critical illnesses were assessed a median of 5 months following discharge and outcomes compared with 100 healthy controls. INTERVENTIONS: None. MEASUREMENTS AND MAIN RESULTS: Parents completed questionnaires documenting child mental and physical well-being, including the Strengths and Difficulties Questionnaires, Chalder Fatigue Scale, and Child Sleep Habits Questionnaire. Children over 8 years completed the Impact of Event Scale -8. The children admitted to PICU scored worse on all measures in comparison with the healthy controls, with 20% scoring at risk for psychiatric disorder, 34% with high levels of post-traumatic stress symptoms, 38% at risk for fatigue disorder, and 80% scoring at risk for sleep disturbance. In the PICU group, multivariable regression analyses identified septic illness as an independent predictor of post-traumatic stress symptoms and family status, past child health problems, and PICU length of stay as predictors of reduced general mental well-being. CONCLUSIONS: Our findings indicate that a significant minority of school-aged children admitted to PICU are at risk for reduced mental and physical well-being in the short term. Symptoms of poor mental well-being were linked to both vulnerability factors and critical illness factors.

Characteristics and outcomes of human parechovirus infection in infants (2008-2012).

UNLABELLED: Human parechoviruses (HPeVs) cause a spectrum of disease ranging from self-limiting illness to severe disease and, sometimes, death. We describe the clinical characteristics and outcomes of HPeV infection in infants. The study describes the clinical and laboratory characteristics and outcomes of infants with HPeV infection during 2008-2012, from three paediatric hospitals in London each with a paediatric intensive care unit. The infants were retrospectively identified through laboratory and patient discharge databases and diagnosed through HPeV PCR. Fifty infants were identified. Half required admission to PICU. Infants less than 3 months were more likely to require PICU (16/25: p < 0.01). Clinical signs at presentation were often indistinguishable from those of bacterial sepsis and meningitis, but inflammatory markers were nearly always (95 % of cases) within normal ranges. Brain MRI showed white matter changes in 10/12 infants. Three of 19 infants with follow-up data (16 %) had significant neurological sequelae. CONCLUSION: HPeV may cause severe disease and long-term neurological sequelae in young infants. HPeV should be considered in infants with clinical features of sepsis/meningitis with normal CSF microscopy. Prospective observational studies are warranted to better define the epidemiology of infection and thus inform future treatment trials.

How in situ simulation affects paediatric nurses' clinical confidence.

Considering the emphasis on safety in health care, new methods for training qualified nurses are being considered. The use of simulation technologies to provide regular and repeated training for qualified nurses in the management of paediatric emergencies has yet to be investigated. This paper presents the results of a study designed to determine if and how a period of regular simulation training in the management of paediatric emergencies improves qualified nurses' clinical confidence. A mixed methods design was employed using a group of qualified paediatric nurses (n=20) who were matched into two groups. The intervention group (n=10) received three simulation-based training sessions and the control group (n=10) had no training. Each nurse completed a pre- and post-clinical confidence questionnaire and were interviewed. Results demonstrated a statistically significant improvement in confidence following simulation training, which was explained by the provision of insight and preparation for real life. Further research should aim to replicate these findings using larger sample sizes and direct assessments of nurses' clinical performance.

Neuropsychologic function three to six months following admission to the PICU with meningoencephalitis, sepsis, and other disorders: a prospective study of school-aged children.

OBJECTIVE: To assess short-term neuropsychological function and academic performance in school children following admission to intensive care and to explore the role of critical neurologic and systemic infection. DESIGN: A prospective observational case-control study. SETTING: Two PICUs. SUBJECTS: A consecutive sample of 88 children aged 5-16 years (median age=10.00, interquartile range=6.00-13.00) who were admitted to intensive care between 2007 and 2010 with meningoencephalitis, septic illness, or other critical illnesses. They were assessed 3 to 6 months following discharge, and their performance was compared with that of 100 healthy controls. Patients were without prior neurologic or neurodevelopmental disorder. INTERVENTIONS: None. MEASUREMENTS AND MAIN RESULTS: Data encompassing demographic and critical illness details were obtained, and children were assessed using tests of intellectual function, memory, and attention. Questionnaires addressing academic performance were returned by teachers. After adjusting for covariates, the children admitted to PICUs significantly underperformed on neuropsychological measures in comparison to healthy controls (p<0.02). Teachers deemed more children admitted to PICUs than controls as performing educationally worse and having problems with school work (ps=0.001), as well as performing below average on aspects of executive function and attention (ps<0.04). Analysis of the effect of illness type on outcome revealed that aspects of neuropsychological function, such as memory function, and teacher-rated academic performance were most reduced in children with meningoencephalitis and septic illness. In the PICU group, multivariable linear regression revealed that worse performance on a composite score of neuropsychologic impairment was more prevalent when children were younger, from a lower social class, and had experienced seizures during their admission (ps<0.02). CONCLUSIONS: Admission to intensive care is followed by deficits in neuropsychologic performance and educational difficulties, with more severe difficulties noted following meningoencephalitis and septic illness. These results highlight the importance of future studies on cognition and educational outcome incorporating type of illness as a moderating factor.

The Imperial Paediatric Emergency Training Toolkit (IPETT) for use in paediatric emergency training: development and evaluation of feasibility and validity.

AIM: To develop and test the feasibility, reliability, and validity of a practical toolkit for the assessment and feedback of skills required to manage paediatric emergencies in critical care settings. METHODS: The Imperial Paediatric Emergency Training Toolkit (IPETT) was developed based on current evidence-base and expert input. IPETT assesses both technical and non-technical skills. The technical component covers skills in the areas of clinical assessment, airway and breathing, cardiovascular, and drugs. The non-technical component is based on the validated NOTECHS tool and covers communication and interaction, cooperation and team skills, leadership and managerial skills, and decision-making. The reliability (internal consistency), content validity (inter-correlations between different skills) and concurrent validity (correlations between global technical and non-technical scores) of IPETT were prospectively evaluated in 45 simulated paediatric crises carried out in a PICU with anaesthetic and paediatric trainees (N=52). Non-parametric analyses were carried out. Significance was set at P<0.05. RESULTS: Cronbach alpha reliability coefficients were overall acceptable for the technical (alpha range=0.638-0.810) and good for the non-technical (alpha range=0.701-0.899) component of IPETT. The median inter-skill correlation was rho=0.564 and rho=0.549 for the technical and non-technical components, respectively. These indicate good content validity, as the skills were inter-related but not redundant. We also demonstrate a correlation between the global technical and non-technical scores (rho=0.471) - all Ps<0.05 during the assessments. CONCLUSION: IPETT offers a psychometrically viable and feasible to use tool in the context of paediatric emergencies training. This study shows that assessment of technical and non-technical skills in combination may offer a more clinically relevant model for training in paediatric emergencies. Further validation should aim to demonstrate skill retention over time and skill transfer from simulation-based training to real emergencies.

The challenge of developing a new predictive formula to estimate energy requirements in ventilated critically ill children.

BACKGROUND: Traditionally, energy requirements have been calculated using predictive equations. These methods have failed to calculate energy expenditure accurately. Routine indirect calorimetry has been suggested, but this method is technically demanding and costly. This study aimed to develop a new predictive equation to estimate energy requirements for critically ill children. METHODS: This prospective, observational study on ventilated children included patients with an endotracheal tube leak of < 10% and fractional inspired oxygen of < 60%. An indirect calorimetry energy expenditure measurement was performed and polynomial regression analysis was used to develop new predictive equations. The new formulas were then compared with existing prediction equations. RESULTS: Data from 369 measurements were included in the formula design. Only weight and diagnosis influenced energy expenditure significantly. Three formulas (A, B, C) with an R² > 0.8 were developed. When we compared the new formulas with commonly used equations (Schofield, Food and Agriculture Organization/World Health Organization/United Nations University, and White equation), all formulas performed very similar, but the Schofield equation seemed to have the lowest SD. CONCLUSIONS: All 3 new pediatric intensive care unit equations have R² values of > 0.8; however, the Schofield equation still performed better than other predictive methods in predicting energy expenditure in these patients. Still, none of the predictive equations, including the new equations, predicted energy expenditure within a clinically accepted range, and further research is required, particularly for patients outside the technical scope of indirect calorimetry.

Isoflurane for life-threatening bronchospasm: a 15-year single-center experience.

BACKGROUND: Children with severe bronchospasm requiring mechanical ventilation may become refractory to conventional therapy. In these critically ill patients, isoflurane is an inhaled anesthetic agent available in some centers to treat bronchospasm. We hypothesized that isoflurane is safe and would lead to improved gas exchange in children with life-threatening bronchospasm refractory to conventional therapy. METHODS: A retrospective review was conducted and included mechanically ventilated children treated with isoflurane in a quaternary pediatric ICU for life-threatening bronchospasm, from 1993 to 2007. Demographic, blood gas, ventilator, and outcome data were collected. RESULTS: Thirty-one patients, with a mean age of 9.5 years (range 0.4-23 years) were treated with isoflurane, from 1993 to 2007. Mean time to initiation of isoflurane after intubation was 13 hours (0-120 h), and the mean maximum isoflurane dose was 1.1% (0.3-2.5%). Mean duration of isoflurane administration was 54.5 hours (range 1-181 h), with a total mean duration of mechanical ventilation of 252 hours (range 16-1,444 h). Isoflurane led to significant improvement in pH and P(CO(2)) within 4 hours of initiation (P ≤ .001). Complications during isoflurane administration included hypotension requiring vasoactive infusions in 24 (77%), arrhythmia in 3 (10%), neurologic side effects in 3 (10%), and pneumothorax in 1 (3%) patient. CONCLUSIONS: Isoflurane led to improvement in pH and P(CO(2)) within 4 hours in this series of mechanically ventilated patients with life-threatening bronchospasm. The majority of patients in this series developed hypotension, but there was a low incidence of other side effects related to isoflurane administration. Isoflurane appears to be an effective therapy in patients with life-threatening bronchospasm refractory to conventional therapy. However, further investigation is warranted, given the uncertain overall impact of isoflurane in this context.

Comparison of pandemic and seasonal influenza reveals higher mortality and increased prevalence of shock in children with severe h1n1/09 infection.

Comparison of the clinical features of H1N1/09 and previous years' influenza A cases reveals that, in children presenting with severe disease, H1N1/09 influenza is associated with an increased prevalence of shock, duration of admission, and mortality. This was not attributable to demographic differences or underlying disease. H1N1/09 influenza is associated with more severe diseases than those with previous years' influenza A strains.

Fulminant pertussis: a multi-center study with new insights into the clinico-pathological mechanisms.

Pertussis carries a high risk of mortality in very young infants. The mechanism of refractory cardio-respiratory failure is complex and not clearly delineated. We aimed to examine the clinico-pathological features and suggest how they may be related to outcome, by multi-center review of clinical records and post-mortem findings of 10 patients with fulminant pertussis (FP). All cases were less than 8 weeks of age, and required ventilation for worsening respiratory symptoms and inotropic support for severe hemodynamic compromise. All died or underwent extra corporeal membrane oxygenation (ECMO) within 1 week. All had increased leukocyte counts (from 54 to 132 x 10(9)/L) with prominent neutrophilia in 9/10. The post-mortem demonstrated necrotizing bronchitis and bronchiolitis with extensive areas of necrosis of the alveolar epithelium. Hyaline membranes were present in those cases with viral co-infection. Pulmonary blood vessels were filled with leukocytes without well-organized thrombi. Immunodepletion of the thymus, spleen, and lymph nodes was a common feature. Other organisms were isolated as follows; 2/10 cases Para influenza type 3, 2/10 Moraxella catarrhalis, 1/10 each with respiratory syncytial virus (RSV), a coliform organism, methicillin-resistant Staphylococcus aureus (MRSA), Haemophilus influenzae, Stenotrophomonas maltophilia, methicillin-sensitive Staphylococcus aureus (MSSA), and candida tropicalis. We postulate that severe hypoxemia and intractable cardiac failure may be due to the effects of pertussis toxin, necrotizing bronchiolitis, extensive damage to the alveolar epithelium, tenacious airway secretions, and possibly leukostasis with activation of the immunological cascade, all contributing to increased pulmonary vascular resistance. Cellular apoptosis appeared to underlay much of these changes. The secondary immuno-compromise may facilitate co-infection.

Successful blind placement of nasojejunal tubes in paediatric intensive care: impact of training and audit.

AIM: This paper is a report of a study to monitor continuing nasojejunal tube placement success rate and to evaluate a training programme for staff placing these tubes at the bedside. BACKGROUND: Gastric delivery of enteral feeds is frequently poorly tolerated due to impaired gastric motility in critically ill children. Consequently, there has been an increased interest in the use of the nasojejunal feeding route. Nasojejunal tubes are both safe and well-tolerated, but placements of these tubes are notoriously difficult and therefore often avoided. Consequently, a blind bedside technique was developed, with a placement success of 96%. METHOD: A training programme using this technique was developed for nursing staff on the unit. This included one-to-one training of key nurses of the nasojejunal tube placement technique and a supervised nasojejunal placement with either the senior nurse or dietitian. This practice was audited in 100 consecutive patients in 2001 and 2004. Patient demographics, diagnosis, time taken for placement and reasons for unsuccessful placements were documented. RESULTS: The 2001 audit (n = 100) indicated that the nasojejunal route was used in 19% of all cases, with 1% and 80% of patients fed via the parenteral and nasogastric route respectively. In 2004 (n = 94), 18% of patients were fed via the nasojejunal route, 3% parenterally and 79% via the gastric route. Placement success continued to be between 94.5-95% in 2001 and 2004. CONCLUSION: Continued successful placement of blind nasojejunal feeding tubes can be achieved through a well-monitored education programme, regular audit cycles and multidisciplinary team support.

Cisatracurium in "weakening doses" assists in weaning from sedation and withdrawal following extended use of inhaled isoflurane.

OBJECTIVE: Isoflurane was used to treat a patient with status asthmaticus refractive to standard therapeutic measures. The patient developed a significant withdrawal syndrome when the isoflurane was weaned. A case is reported here where this withdrawal syndrome was treated successfully by using a weakening dose neuromuscular blockade with cisatracurium. DESIGN: Case report. SETTING: Pediatric critical care unit. PATIENT: A 4-yr-old girl with severe reactive airways disease. INTERVENTIONS: The use of weakening doses of cisatracurium to assist in weaning from mechanical ventilation in the setting of withdrawal symptoms following the extended use of inhaled isoflurane. MEASUREMENTS AND MAIN RESULTS: Despite treatment with mechanical ventilation, intravenous corticosteroids, and bronchodilators for status asthmaticus, the patient required inhaled isoflurane. She became tolerant to isoflurane over an extended period of time; her tolerance was associated with a specific withdrawal syndrome, with the development of choreoathetoid movements resulting in poor pulmonary coordination and agitation. Conventional medical treatment of withdrawal failed. Finally, by using an infusion of cisatracurium at weakening doses to assist in the control of these choreoathetoid movements, the isoflurane and ventilator support were weaned. CONCLUSIONS: Weakening doses of cisatracurium may be used safely to control unpleasant motor symptoms secondary to tolerance of isoflurane. This may have a use in other circumstances where agitation in mechanically ventilated patients is not due to pain or anxiety.

Arginine and nitric oxide metabolism in critically ill septic pediatric patients.

OBJECTIVE: To investigate whole body, arginine metabolism and nitric oxide synthesis rates in septic, critically ill pediatric patients. DESIGN: Prospective study. SETTING: Pediatric intensive care unit at a general hospital. PATIENTS: Ten consecutive septic patients age 6-16 yrs. INTERVENTIONS: Septic patients received an 8-hr primed, constant intravenous tracer infusion of L-[guanidino-15N2]arginine, L-[1-13C]leucine, and [13C]urea. A 24-hr urine collection was obtained for determination of [15N]nitrate enrichment (15NO3(-)) and urinary nitrogen. The next day they received an infusion of L-[5-13C]arginine and L-[5-13C-ureido, 5,5, 2H2]citrulline. Blood samples were obtained for determination of plasma isotopic enrichment of the tracers given and of derived [15N]citrulline (nitric oxide synthesis), L-[13C-guanidino 5,5, 2H2]arginine (M+3 arg) (arginine synthesis), and [15N]urea (urea formation). Data are compared with historic controls from studies in healthy young adults. MEASUREMENTS AND MAIN RESULTS: Plasma arginine fluxes were 67 +/- 21 and 72 +/- 17 micromol x kg(-1) x hr(-1), respectively, for the [15N2 guanidino] and the [13C] arginine labels, which were not different from reported adult values. The rates of arginine oxidation were 22.9 +/- 10.8 micromol x kg(-1) x hr(-1) and were higher than arginine synthesis rates of 9.6 +/- 4.2 micromol x kg(-1) x hr(-1) (p <.01); therefore, these patients were in a negative arginine balance. The rates of nitric oxide synthesis as estimated by the [15N]citrulline method were 1.58 +/- 0.69 micromol x kg(-1) x hr(-1) for septic patients and higher (p <.05) than values of 0.96 +/- 0.1 micromol x kg(-1) x hr(-1) in healthy adults. Septic patients were in a negative protein (leucine) balance of about -1.00 +/- 0.40 g x kg(-1) x day(-1). CONCLUSIONS: Homeostasis of plasma arginine in septic patients was impaired compared with reported adult values. The rates of arginine oxidation were increased whereas net arginine synthesis was unchanged, leading to a negative arginine balance. The rates of nitric oxide synthesis and the fraction of plasma arginine used for nitric oxide and urea formation were increased. These findings suggest that under condition of sepsis, arginine becomes essential in critically ill children.