Pediatric Critical Care Medicine

Epidemiology of Pediatric Severe Sepsis in Main PICU Centers in Southwest China Objectives: To estimate the prevalence, management, and outcomes of pediatric severe sepsis in the main PICUs in Southwest China. Design: A prospective, observational, and multicenter study. Setting: Eight PICUs in Southwest China with 19 (13–24) beds and 1,322 (1,066–1,452) annual admissions each. Patients: A total of 10,598 patients (29 d to 18 yr old) were consecutively admitted between September 1, 2016, and August 31, 2017. All patients were screened and evaluated for severe sepsis or septic shock. Of them, 10,353 patients were excluded due to incomplete data or not meeting the consensus criteria for severe sepsis or septic shock; 245 patients were included with complete data. Interventions: None. Measurements and Main Results: Finally, 245 patients who were diagnosed with severe sepsis or septic shock were included in the study, with an incidence rate of 2.3%. Of them, 64.0% of the enrolled patients were male with 80.8% being less than 5 years old and 60.8% being from rural areas. The respiratory system was the most common organ system in which dysfunction was observed (76.7%) as well as the most frequently infected site (37.6%). The primary therapies were antibiotics (99.0%), immunoglobulin (88.3%), mechanical ventilation (78.4%), vasoactive infusions (59.6%), and corticosteroids (46.1%). Among the 188 patients who had respiratory dysfunction, 173(92%) required mechanical ventilation and 39 (20.7%) met the criteria for pediatric acute respiratory distress syndrome. Seven of the patients with pediatric acute respiratory distress syndrome died (7/39, 17.9%). The median durations for mechanical ventilation and vasoactive medications were 123.5 hours (35.25–226.00 hr) and 2 days (1–5 d), respectively. Eighty-six percent of patients had multiple organ dysfunction syndrome at the point at which severe sepsis was recognized, and 31% had underlying conditions. The hospital mortality rate was 18.8%. Conclusions: This report is the first to present the prevalence, treatment, and outcomes of pediatric severe sepsis in the main PICU centers in Southwest China. The mortality rate remains high; therefore, improved clinical management and implementation of large-scale clinical trials are necessary to improve early diagnoses and treatment. Dr. Xu received support for article research from Chongqing Science and Technology Bureau. Drs. C. Xiao, S. Wang, Fang, Xu, S. Xiao, Li, Zhang, Luo, Jiang, Huang, Y. Chen, J. Chen, Yu, D. Ren, X. Ren, and Tang disclosed government work. Dr. H. Wang has disclosed that he does not have any potential conflicts of interest. For information regarding this article, E-mail: xufeng9899@163.com ©2019The Society of Critical Care Medicine and the World Federation of Pediatric Intensive and Critical Care Societies

Barriers to Communication in a PICU: A Qualitative Investigation of Family and Provider Perceptions Objectives: Family and medical provider perceptions of communication barriers within the PICU are poorly understood. We designed a qualitative study to characterize the perspective of families and medical providers of critically ill children regarding communication barriers. The identified barriers may be used to direct efforts to improve communication. Design: Semi-structured interviews were conducted from August 2017 to January 2018. Interviews were audio recorded and professionally transcribed verbatim. Setting: A PICU at a tertiary care academic center. Patients: Forty-two families whose children were admitted to the PICU (excluding patients receiving end-of-life care or with protective services involvement) and 12 PICU staff members, including nurses, residents, fellows, and attending's. Interventions: None. Measurements and Main Results: An interprofessional team of a physician, nurse manager, and medical student coded the transcripts. Interviewing continued until thematic saturation was reached. Codes were organized into common themes using a modified constant comparative method. The families interviewed represented 16 previously healthy children, and 26 children with a chronic health condition. Staff interviewed included three residents, three fellows, three attending intensivists, and three nurses. Participants' perceptions and experiences of barriers to communication included the following: 1) Communication breakdowns related to coordination of care among several services, 2) Family-centered rounds are insufficient for effective communication, 3) Undervaluing the knowledge of families of children with chronic health conditions or special needs, and 4) Communication breakdowns occur across provider hand-offs. Theme 3 was identified by families, but not by providers. Conclusions: Families and medical providers both identified several barriers to communication. However, only families identified the barrier "Undervaluing the knowledge of families with chronically ill children." Future work should explore these barriers and the discrepancy in perception between providers and families to determine if there are interventions that improve both family satisfaction and patient care. Supplemental digital content is available for this article. Direct URL citations appear in the printed text and are provided in the HTML and PDF versions of this article on the journal's website (http:/journals.lww.com/pccmjournal). Supported, in part, by grant from the Yale School of Medicine Office of Student Research and the Yale School of Medicine Department of Pediatrics, Section of Critical Care. Mr. Greenway received funding from Yale School of Medicine, Office of Student Research and Yale School of Medicine, Department of Pediatrics, Section of Critical Care. The remaining authors have disclosed that they do not have any potential conflicts of interest. Address requests for reprints to: Kevin G. Couloures, DO, MPH, Department of Pediatrics, 333 Cedar Street, New Haven, CT 06520-8064. E-mail: kevin.couloures@yale.edu ©2019The Society of Critical Care Medicine and the World Federation of Pediatric Intensive and Critical Care Societies

Sedative and Analgesic Drug Rotation Protocol in Critically Ill Children With Prolonged Sedation. Evaluation of Implementation and Efficacy to Reduce Withdrawal Syndrome Objectives: The first aim of this study was to assess the implementation of a sedative and analgesic drug rotation protocol in a PICU. The second aim was to analyze the incidence of withdrawal syndrome, drug doses, and time of sedative or analgesic drug infusion in children after the implementation of the new protocol. Design: Prospective observational study. Setting: PICU of a tertiary care hospital between June 2012 and June 2016. Patients: All patients between 1 month and 16 years old admitted to the PICU who received continuous IV infusion of sedative or analgesic drugs for more than 4 days were included in the study. Interventions: A sedative and analgesic drug rotation protocol was designed. The level of sedation, analgesia, and withdrawal syndrome were monitored with validated scales. The relationship between compliance with the protocol and the incidence of withdrawal syndrome was studied. Measurements and Main Results: One-hundred pediatric patients were included in the study. The protocol was followed properly in 35% of patients. Sixty-seven percent of the overall cohort presented with withdrawal syndrome. There was a lower incidence rate of withdrawal syndrome (34.3% vs 84.6%; p < 0.001), shorter PICU length of stay (median 16 vs 25 d; p = 0.003), less time of opioid infusion (median 5 vs 7 d for fentanyl; p = 0.004), benzodiazepines (median 5 vs 9 d; p = 0.001), and propofol (median 4 vs 8 d; p = 0.001) in the cohort of children in which the protocol was followed correctly. Conclusions: Our results show that compliance with the drug rotation protocol in critically ill children requiring prolonged sedation may reduce the appearance of withdrawal syndrome without increasing the risk of adverse effects. Furthermore, it may reduce the time of continuous IV infusions for most sedative and analgesic drugs and the length of stay in PICU. Dr. Sanavia involved in patient recruitment, data collection, analysis and interpretation of data, and writing up of the article. Dr. Mencía involved in study design, analysis and interpretation of data, and drafting the article or revising it critically for important intellectual content. Dr. Lafever involved in data collection and translation. Dr. Solana involved in patient recruitment and data collection. Dr. Garcia involved in data collection. Dr. López-Herce involved in drafting the article or revising it critically for important intellectual content and final approval of the version to be published. Supported, in part, by Dodot Scholarship from the Spanish Pediatric Association. Maternal and Child Health and Development Network. RETICS Health Research Networks financed by the R + D + I PN 2008–2011, ISCIII (Health Institute Carlos III)—General Subdirectorate for Evaluation and Promotion of Research and the European Regional Development Fund, ef. RD16/0022/0007. The authors have disclosed that they do not have any potential conflicts of interest. For information regarding this article, E-mail: eva.sanavia@hotmail.com ©2019The Society of Critical Care Medicine and the World Federation of Pediatric Intensive and Critical Care Societies

Critical Trio Exome Benefits In-Time Decision-Making for Pediatric Patients With Severe Illnesses Objectives: Critical illnesses caused by undiagnosed genetic conditions are challenging in PICUs. Whole-exome sequencing is a powerful diagnostic tool but usually costly and often fail to arrive at a final diagnosis in a short period. We assessed the feasibility of our whole-exome sequencing as a tool to improve the efficacy of rare diseases diagnosis for pediatric patients with severe illness. Design: Observational analysis. Method: We employed a fast but standard whole-exome sequencing platform together with text mining-assisted variant prioritization in PICU setting over a 1-year period. Setting: A tertiary referral Children's Hospital in Taiwan. Patients: Critically ill PICU patients suspected of having a genetic disease and newborns who were suspected of having a serious genetic disease after newborn screening were enrolled. Interventions: None. Measurements and Main Results: Around 50,000 to 100,000 variants were obtained for each of the 40 patients in 5 days after blood sampling. Eleven patients were immediately found be affected by previously reported mutations after searching mutation databases. Another seven patients had a diagnosis among the top five in a list ranked by text mining. As a whole, 21 patients (52.5%) obtained a diagnosis in 6.2 ± 1.1 working days (range, 4.3–9 d). Most of the diagnoses were first recognized in Taiwan. Specific medications were recommended for 10 patients (10/21, 47.6%), transplantation was advised for five, and hospice care was suggested for two patients. Overall, clinical management was altered in time for 81.0% of patients who had a molecular diagnosis. Conclusions: The current whole-exome sequencing algorithm, balanced in cost and speed, uncovers genetic conditions in infants and children in PICU, which helps their managements in time and promotes better utilization of PICU resources. Supplemental digital content is available for this article. Direct URL citations appear in the printed text and are provided in the HTML and PDF versions of this article on the journal's website (http://journals.lww.com/pccmjournal). Supported, in part, by the Ministry of Science and Technology, R.O.C., under grant number MOST 106-3114-B-002-009. Dr. Hwu's institution received funding from Ministry of Science and Technology. Miss. Chen disclosed work for hire. The remaining authors have disclosed that they do not have any potential conflicts of interest. For information regarding this article, E-mail: ncleentu@ntu.edu.tw ©2019The Society of Critical Care Medicine and the World Federation of Pediatric Intensive and Critical Care Societies

Epidemiology of Clinically Relevant Bleeding in Critically Ill Adolescents Objectives: The epidemiology of clinically relevant bleeding in critically ill adolescents, particularly those who are at high risk of venous thromboembolism, is unclear. In preparation for a randomized clinical trial of pharmacologic prophylaxis against venous thromboembolism, we characterized the epidemiology of clinically relevant bleeding in critically ill adolescents. Design: Post hoc analysis of data from a pediatric multicenter observational study of venous thromboembolism. Setting: Six PICUs. Patients: Adolescents 13–17 years old who received cardiac or pulmonary support for at least 48 hours were eligible. Those admitted with venous thromboembolism or receiving therapeutic anticoagulation were excluded. Interventions: None. Measurements and Main Results: Adolescents (n = 88) were followed daily for the development of any bleeding event. The severity of the event was categorized based on the definitions by the International Society on Thrombosis and Haemostasis. The frequency of clinically relevant bleeding was 29.5% (95% CI, 20.3–40.2%) or 3.7 events (95% CI, 2.5–5.4 events) per 100 patient-days. Adolescents with venous thromboembolism were more likely to develop clinically relevant bleeding (hazard ratio, 2.06; 95% CI, 1.08–3.94). Age was negatively associated with clinically relevant bleeding (hazard ratio for every 1-year increase in age: 0.68; 95% CI, 0.58–0.79). In contrast, predicted risk of mortality (hazard ratio for every 0.10 increase in risk: 1.35; 95% CI, 1.05–1.74) and admission for trauma or surgery (hazard ratio: 2.04; 95% CI, 1.21–3.44) were positively associated with clinically relevant bleeding. The association of clinically relevant bleeding with medications, interventions, or laboratory tests, including mechanical ventilation and pharmacologic prophylaxis with anticoagulation, did not reach statistical significance. Adolescents with clinically relevant bleeding stayed in the hospital longer than those without clinically relevant bleeding. Conclusions: Clinically relevant bleeding is common in critically ill adolescents who are at high risk of venous thromboembolism. Admission for trauma or surgery can be used to stratify the risk of clinically relevant bleeding in these adolescents. Supported, in part, by grant from the American Heart Association Award Number 14CRP20490002 to Dr. Faustino. It was also made possible by Clinical and Translational Science Award grant number UL1 RR024139 from the National Center for Research Resources and the National Center for Advancing Translational Science, components of the National Institutes of Health (NIH), and NIH roadmap for Medical Research to Dr. Shabanova, and an unrestricted grant support from Layla's Dream to Dr. Polikoff. For information regarding this article, E-mail: Matthew.Pinto@wmchealth.org ©2019The Society of Critical Care Medicine and the World Federation of Pediatric Intensive and Critical Care Societies

Functional Status Scale: Cross-Cultural Adaptation and Validation in Brazil Objective: The goal of the present study was to perform a cross-cultural adaptation and clinical validation of the Functional Status Scale for use in the Brazilian population. Design: Cross-cultural adaptation study followed by a cross-sectional validation study. Setting: Single-center PICU at a hospital in Porto Alegre, Brazil. Patients: Children and adolescents of both sexes, 1 month and under 18 years old, who had been treated at the PICU. Interventions: The cross-cultural adaptation consisted of the following stages: translation, synthesis of the translated versions, back translations, synthesis of the back translations, committee review, and pretesting. For the clinical validation stage, the Brazilian Functional Status Scale was applied within 48 hours after discharge from the PICU. The Brazilian Functional Status Scale's reliability and validity properties were tested. Measurements and Main Results: A total of 314 patients were evaluated. Median age was 24 months (7.0–105.0 mo), 54.1% were males, and their overall functional score was 9 ± 2.8. The Brazilian Functional Status Scale demonstrated excellent interobserver reliability, with an intraclass correlation coefficient of 0.98, and κ coefficients between 0.716 and 1.000 for the functional domains, which indicated good to excellent agreement. Using the Bland-Altman method, we confirmed low variability among the evaluator's responses (0.93 to –1.06 points). Regarding the Brazilian Functional Status Scale's content validity, there was a correlation between length of PICU stay (r = 0.378; p < 0.001) and time on invasive mechanical ventilation (r = 0.261; p < 0.05), and the test could discriminate between groups with different comorbidity levels (p < 0.001). Conclusions: The Functional Status Scale has been culturally adapted and validated for use in Brazil and is now available for use in the assessment of functionality in Brazilian children and adolescents. Supplemental digital content is available for this article. Direct URL citations appear in the printed text and are provided in the HTML and PDF versions of this article on the journal's website (http://journals.lww.com/pccmjournal). Supported, in part, by grants from the Higher Education Improvement Coordination (CAPES). The authors have disclosed that they do not have any potential conflicts of interest. This work was performed at the Hospital da Criança Santo Antônio, Irmandade Santa Casa de Misericórdia de Porto Alegre, Porto Alegre (RS), Brazil. For information regarding this article, E-mail: gabi.apereira@gmail.com ©2019The Society of Critical Care Medicine and the World Federation of Pediatric Intensive and Critical Care Societies

The Base Deficit, International Normalized Ratio, and Glasgow Coma Scale (BIG) Score, and Functional Outcome at Hospital Discharge in Children With Traumatic Brain Injury Objectives: To examine the association of the base deficit, international normalized ratio, and Glasgow Coma Scale (BIG) score on emergency department arrival with functional dependence at hospital discharge (Pediatric Cerebral Performance Category ≥ 4) in pediatric multiple trauma patients with traumatic brain injury. Design: A retrospective cohort study of a pediatric trauma database from 2001 to 2018. Setting: Level 1 trauma program at a university-affiliated pediatric institution. Patients: Two to 17 years old children sustaining major blunt trauma including a traumatic brain injury and meeting trauma team activation criteria. Interventions: None. Measurements and Main Results: Two investigators, blinded to the BIG score, determined discharge Pediatric Cerebral Performance Category scores. The BIG score was measured on emergency department arrival. The 609 study patients were 9.7 ± 4.4 years old with a median Injury Severity Score 22 (interquartile range, 12). One-hundred seventy-one of 609 (28%) had Pediatric Cerebral Performance Category greater than or equal to 4 (primary outcome). The BIG constituted a multivariable predictor of Pediatric Cerebral Performance Category greater than or equal to 4 (odds ratio, 2.39; 95% CI, 1.81–3.15) after adjustment for neurosurgery requirement (odds ratio, 2.83; 95% CI, 1.69–4.74), pupils fixed and dilated (odds ratio, 3.1; 95% CI, 1.49–6.38), and intubation at the scene or referral hospital (odds ratio, 2.82; 95% CI, 1.35–5.87) and other postulated predictors of poor outcome. The area under the BIG receiver operating characteristic curve was 0.87 (0.84–0.90). Using an optimal BIG cutoff less than or equal to 8, sensitivity and negative predictive value for functional dependence at discharge were 93% and 96%, respectively, compared with a sensitivity of 79% and negative predictive value of 91% with Glasgow Coma Scale less than or equal to 8. In children with Glasgow Coma Scale 3, the BIG score was associated with brain death (odds ratio, 2.13; 95% CI, 1.58–2.36). The BIG also predicted disposition to inpatient rehabilitation (odds ratio, 2.26; 95% CI, 2.17–2.35). Conclusions: The BIG score is a simple, rapidly obtainable severity of illness score that constitutes an independent predictor of functional dependence at hospital discharge in pediatric trauma patients with traumatic brain injury. The base deficit, international normalized ratio, and Glasgow Coma Scale score may benefit Trauma and Neurocritical care programs in identifying ideal candidates for traumatic brain injury trials within the therapeutic window of treatment. Presented, in part, as oral presentations at the: 1) Trauma Association of Canada meeting, Toronto, ON, Canada, February 23, 2018; and 2) Pediatric Academic Societies meeting, Toronto, ON, Canada, May 6, 2018. Drs. Davis, Hochstadter, and Schuh conceived the study and wrote the protocol. Dr. Davis supervised data collection by Drs. Hochstadter and Daya. Mr. Stephens performed the data analysis. Drs. Davis, Hochstadter, and Schuh drafted the article and Drs. Davis, Hochstadter, Daya, Kulkarni, Wales, and Schuh contributed substantially to its revision. All authors take responsibility for the article as a whole. The authors have disclosed that they do not have any potential conflicts of interest. Address requests for reprints to: Adrienne Davis, MD, MSc, FRCPC, Division of Pediatric Emergency Medicine, Department of Pediatrics, Hospital for Sick Children, University of Toronto, Toronto, ON, Canada. E-mail: adrienne.davis@sickkids.ca ©2019The Society of Critical Care Medicine and the World Federation of Pediatric Intensive and Critical Care Societies

Endothelial-Dependent Vasomotor Dysfunction in Infants After Cardiopulmonary Bypass Objectives: Cardiopulmonary bypass–induced endothelial dysfunction has been inferred by changes in pulmonary vascular resistance, alterations in circulating biomarkers, and postoperative capillary leak. Endothelial-dependent vasomotor dysfunction of the systemic vasculature has never been quantified in this setting. The objective of the present study was to quantify acute effects of cardiopulmonary bypass on endothelial vasomotor control and attempt to correlate these effects with postoperative cytokines, tissue edema, and clinical outcomes in infants. Design: Single-center prospective observational cohort pilot study. Setting: Pediatric cardiac ICU at a tertiary children's hospital. Patients: Children less than 1 year old requiring cardiopulmonary bypass for repair of a congenital heart lesion. Intervention: None. Measurements and Main Results: Laser Doppler perfusion monitoring was coupled with local iontophoresis of acetylcholine (endothelium-dependent vasodilator) or sodium nitroprusside (endothelium-independent vasodilator) to quantify endothelial-dependent vasomotor function in the cutaneous microcirculation. Measurements were obtained preoperatively, 2–4 hours, and 24 hours after separation from cardiopulmonary bypass. Fifteen patients completed all laser Doppler perfusion monitor (Perimed, Järfälla, Sweden) measurements. Comparing prebypass with 2–4 hours postbypass responses, there was a decrease in both peak perfusion (p = 0.0006) and area under the dose-response curve (p = 0.005) following acetylcholine, but no change in responses to sodium nitroprusside. Twenty-four hours after bypass responsiveness to acetylcholine improved, but typically remained depressed from baseline. Conserved endothelial function was associated with higher urine output during the first 48 postoperative hours (R2 = 0.43; p = 0.008). Conclusions: Cutaneous endothelial dysfunction is present in infants immediately following cardiopulmonary bypass and recovers significantly in some patients within 24 hours postoperatively. Confirmation of an association between persistent endothelial-dependent vasomotor dysfunction and decreased urine output could have important clinical implications. Ongoing research will explore the pattern of endothelial-dependent vasomotor dysfunction after cardiopulmonary bypass and its relationship with biochemical markers of inflammation and clinical outcomes. The views expressed in this article are those of the author(s) and do not necessarily reflect the official policy or position of the Department of the Navy, Department of Defense, or the U.S. Government. Dr. Krispinsky is a military service member. This work was prepared as part of his official duties. Title 17 U.S.C. 105 provides that "Copyright protection under this title is not available for any work of the U.S. Government." Title 17 U.S.C. 101 defines a U.S. Government work as a work prepared by a military service member or employee of the U.S. Government as part of that person's official duties. Supplemental digital content is available for this article. Direct URL citations appear in the printed text and are provided in the HTML and PDF versions of this article on the journal's website (http://journals.lww.com/pccmjournal). Supported, in part, by grants from Vanderbilt University Medical Center. Dr. Stark is supported by a grant from the National Institute of General Medical Sciences (K08 GM117367). Dr. Krispinsky disclosed government work. Dr. Stark's institution received funding from the National Institutes of Health (NIH), and he received support for article research from the NIH. The remaining authors have disclosed that they do not have any potential conflicts of interest. For information regarding this article, E-mail: luke.t.krispinsky.mil@mail.mil Address requests for reprints to: Fred S. Lamb, MD, PhD, Pediatric Critical Care, Vanderbilt Children's Hospital, 2200 Children's Way, 5111 Doctors' Office Tower, Nashville, TN 37232. E-mail: fred.s.lamb@vanderbilt.edu ©2019The Society of Critical Care Medicine and the World Federation of Pediatric Intensive and Critical Care Societies

Current Epidemiology of Vocal Cord Dysfunction After Congenital Heart Surgery in Young Infants Objectives: Surgery of the aortic arch poses risk of recurrent laryngeal nerve injury due to the anatomic proximity and can manifest as vocal cord dysfunction after surgery. We assessed risk factors for vocal cord dysfunction and calculated surgical procedure associated rates in young infants after congenital heart surgery. Design: Cross section analysis. Setting: Forty-four children's hospitals reporting administrative data to Pediatric Health Information System. Participants: Cardiac surgical patients less than or equal to 90 days old and discharged between January 2004 and June 2014. Interventions: None. Measurements and Main Results: Overall, 2,319 of 46,567 subjects (5%) had vocal cord dysfunction, increasing from 4% to 7% over the study period. Of those with vocal cord dysfunction, 75% had unilateral partial paralysis. Vocal cord dysfunction was significantly more common in newborn infants (74%), those with aortic arch procedures (77%) and with greater surgical complexity. Rates of vocal cord dysfunction ranged from 0.7% to 22.4% across surgical procedure groups. Vocal cord dysfunction was significantly associated with greater use of: prolonged mechanical ventilation (53% vs 40%), diaphragmatic plication (3% vs 1%), feeding tube use (32% vs 8%), surgical airways (4% vs 2%), and prolonged length of stay (44 vs 21 d). Vocal cord dysfunction testing increased significantly over the study (6–14 %), and vocal cord dysfunction diagnosis increased almost two-fold (odds ratio, 1.9; 95% CI, 1.7–2.1) comparing the last to first study quarters with the increase in vocal cord dysfunction diagnosis occurring predominately in surgeries to the aortic arch supported by cardiopulmonary bypass. However, aortic procedures without cardiopulmonary bypass and nonaortic arch procedures were common surgeries accounting for 27% and 23% of vocal cord dysfunction cases despite low overall vocal cord dysfunction rates (3.7% and 2.6%). Conclusions: Vocal cord dysfunction complicated all cardiac surgical procedures among infants including those without aortic arch involvement. Increased efforts to determine appropriate indications for prevention, screening and treatment of vocal cord dysfunction among young infants after congenital heart surgery are needed. This work was performed at the University of Utah. Supplemental digital content is available for this article. Direct URL citations appear in the printed text and are provided in the HTML and PDF versions of this article on the journal's website (http:/journals.lww.com/pccmjournal). The authors have disclosed that they do not have any potential conflicts of interest. For information regarding this article, E-mail: Susan.Bratton@hsc.utah.edu ©2019The Society of Critical Care Medicine and the World Federation of Pediatric Intensive and Critical Care Societies

Anemia at Discharge From PICU: A Bicenter Descriptive Study Objective: To determine the prevalence and risk markers of anemia at PICU discharge. Design: Bicenter retrospective cohort study. Setting: Two multidisciplinary French PICUs. Patients: All children admitted during a 5-year period, staying in the PICU for at least 2 days, and for whom a hemoglobin was available at PICU discharge. Interventions: None. Measurements and Main Results: Patient, admission, and PICU stay characteristics were retrospectively collected in the electronic medical records of each participating PICU. Anemia was defined according to the World Health Organization criteria. Among the 3,170 patients included for analysis, 1,868 (58.9%) were anemic at discharge from PICU. The proportion of anemic children differed between age categories, whereas the median hemoglobin level did not exhibit significant variations according to age. After multivariate adjustment, anemia at PICU admission was the strongest predictor of anemia at PICU discharge, and the strength of this association varied according to age (interaction). Children anemic at PICU admission had a reduced risk of anemia at PICU discharge if transfused with RBCs during the PICU stay, if less than 6 months old, or if creatinine level at PICU admission was low. Children not anemic at PICU admission had an increased risk of anemia at PICU discharge if they were thrombocytopenic at PICU admission, if they had higher C-reactive protein levels, and if they received plasma transfusion, inotropic/vasopressor support, or mechanical ventilation during the PICU stay. Conclusions: Anemia is frequent after pediatric critical illness. Anemia status at PICU admission defines different subgroups of critically ill children with specific prevalence and risk markers of anemia at PICU discharge. Further studies are required to confirm our results, to better define anemia during pediatric critical illness, and to highlight the causes of post-PICU stay anemia, its course, and its association with post-PICU outcomes. This work was performed at the Hôpital Universitaire Jeanne de Flandre, Service de Réanimation Pédiatrique 2, Avenue Eugène Avinée, 59037 Lille Cedex, France, and Hôpital Universitaire Femme Mère Enfant, Hospices Civils de Lyon, Service de Réanimation Pédiatrique, 59, Boulevard Pinel, 69677 Lyon-Bron, France. Supplemental digital content is available for this article. Direct URL citations appear in the printed text and are provided in the HTML and PDF versions of this article on the journal's website (http://journals.lww.com/pccmjournal). The authors have disclosed that they do not have any potential conflicts of interest. For information regarding this article, E-mail: pierre.demaret@chc.be ©2019The Society of Critical Care Medicine and the World Federation of Pediatric Intensive and Critical Care Societies

Alexandros Sfakianakis
Anapafseos 5 . Agios Nikolaos
Crete.Greece.72100
2841026182

6948891480

alsfakia@gmail.com