Characterization of pediatric patients receiving prolonged mechanical ventilation

Transcription

1 Characterization of ediatric atients receiving rolonged mechanical ventilation Ezequiel Monteverde, MD; Analía Fernández, MD; Rossana Poterala, MD; Nilda Vidal, MD; Alejandro Siaba Serrate, MD; Pablo Castelani, MD; Lidia Albano, MD; Fernanda Podestá, MD; Julio A. Farias, MD Objective: To describe the characteristics and risk factors of ediatric atients who receive rolonged mechanical ventilation, defined as ventilatory suort for >21 days. Design: Prosective cohort. Setting: Four medical-surgical ediatric intensive care units in four university-affiliated hositals in Argentina. Patients: All consecutive atients from 1 month to 15 yrs old admitted to articiating ediatric intensive care units from June 1, 2007, to August 31, 2007, who received mechanical ventilation (invasive or noninvasive) for >12 hrs. Interventions: None. Measurements and Main Results: Demograhic and hysiologic data on admission to the ediatric intensive care units, drugs and events during the study eriod, and outcomes were rosectively recorded. A total of 256 atients were included. Of these, 23 (9%) required mechanical ventilation for >21 days and were assigned to the rolonged mechanical ventilation grou. Patients requiring rolonged mechanical ventilation had higher mortality (43% vs. 21%, <.05) and longer ediatric intensive care unit stay 35 days [28 64 days] vs. 10 days [6 14]). There was no difference between the grous in age and gender distribution, reasons for admission, incidence of immunodeficiencies, or Paediatric Index of Mortality 2 score. The only difference at admission was a higher rate of genetic diseases in rolonged mechanical ventilation atients (26% vs. 9%, <.05). There was a higher incidence of setic shock (87% vs. 34%, <.01), acute resiratory distress syndrome (43% vs. 20%, <.01), and ventilator-associated neumonia (35% vs. 8%, <.01) and higher utilization of doamine (78% vs. 42%, <.01), noreinehrine (61% vs. 15%, <.01), multile antibiotics (83% vs. 20%, <.01), and blood transfusions (52% vs. 14%, <.01). The roortion of extubation failure was higher in the rolonged mechanical ventilation grou with similar rates of unlanned extubations in both grous. Variables remaining significantly associated with rolonged mechanical ventilation after multivariate analysis were treatment with multile antibiotics, setic shock, ventilator-associated neumonia, and use of noreinehrine. Conclusions: Patients with rolonged mechanical ventilation have more comlications and require more ediatric intensive care unit resources. Mortality in these atients dulicates that from those requiring shorter suort. (Pediatr Crit Care Med 2011; 12: ) KEY WORDS: mechanical ventilation; mortality; ediatric intensive care unit; setic shock; ventilator-associated neumonia From the Unidad de Cuidados Intensivos (EM, AF, RP, JF), Hosital de Niños Ricardo Gutiérrez, Ciudad Autónoma de Buenos Aires; Unidad de Cuidados Intensivos Pediátricos (NV, LA), Hosital Nacional Alejandro Posadas, Haedo, Buenos Aires; Unidad de Cuidados Intensivos Pediátricos (ASS), Hosital Austral, Pilar, Buenos Aires; and Unidad de Cuidados Intensivos (PC, FP), Hosital de Niños Sor María Ludovica, La Plata, Buenos Aires, Argentina. The authors have not disclosed any otential conflicts of interest. For information regarding this article, jufarias@intramed.net Coyright 2011 by the Society of Critical Care Medicine and the World Federation of Pediatric Intensive and Critical Care Societies DOI: /PCC.0b013e c0b Advances in ediatric intensive care have made ossible a new and emerging grou of atients: those who survive thanks to imrovements in mechanical ventilation (MV) rotocols and in the early detection and treatment of setic shock, but who suffer chronic critical illness with a need for rolonged mechanical ventilation (PMV) (1). PMV has been defined by the National Association for Medical Direction on Resiratory Care as the need for at least 6 hrs of ventilatory suort er day for 21 consecutive days (2). In adult oulations PMV is well known to be associated with longer hosital stays and also with higher morbidity and mortality. While these associations have also been observed in a ediatric oulation (3), it is still not clear that children who require PMV have the same characteristics with the same risk factors. The object of this study is to resent a descrition of the characteristics and risk factors of children who received PMV in four third-level ediatric intensive care units (PICUs) in Argentina. Such a descrition should be useful in develoing secific interventions that might imrove the quality of care such children receive. MATERIALS AND METHODS Study Design. The investigation was a rosective cohort study carried out in four medical-surgical PICUs at four tertiary level university-affiliated hositals. The four PICUs selected for this study were regional referral centers in 1) the rovince of Buenos Aires, to wit, Hosital de Niños Sor María Ludovica (La Plata, 20 PICU beds), Hosital Nacional Alejandro Posadas (Haedo, 20 PICU beds), and Hosital Universitario Austral (Pilar, nine PICU beds), and 2) the City of Buenos Aires, namely, Hosital de Niños Ricardo Gutiérrez (26 PICU beds). All units admitted both clinical and surgical atients. The study was aroved by the Institutional Review Board of each institution, and informed consent was waived since this was an observational study with no interventions. Patients. All atients from 1 month to 15 yrs old who were admitted to the PICUs between June 1 and August 31, 2007, inclusive, who received MV (invasive or noninvasive) for 12 hrs were included. Patients already receiving chronic ventilatory suort were excluded. 1

2 Study Protocol. Two investigators at each unit were trained to rosectively collect all information required by the study rotocol and enter it on a secially designed form. Since these were the only hysicians who were aware of the study, erformance in the PICUs was not affected. Study variables with a robable relationshi to PMV were screened daily until the end of MV or until day 21, whichever was first. Afterward, follow-u was only carried out to record the duration of MV, the length of stay at the PICU, and the date of death of those atients who died. Data recorded at admission were the date of admission to the hosital and to the PICU, age, gender, weight, chronic comlex conditions (4), rimary source of admission (emergency room, hosital ward, other hosital), severity of illness and redicted mortality at admission (Paediatric Index of Mortality 2 score), tye of admission (medical or surgical), and rimary reason for initiation of MV, divided into four categories: acute on chronic resiratory disease, acute resiratory failure, altered mental status, and neuromuscular disease. The first category included atients with a diagnosis of chronic lung disease (resiratory malformations, chronic obstructive ulmonary disease, and cystic fibrosis or asthma) with exacerbation requiring MV. The second included acute viral bronchiolitis, neumonia (community or hosital acquired), acute resiratory distress syndrome (ARDS), uer airway obstruction, acute ulmonary edema and/or congestive heart failure, ostoerative state, asiration, trauma, cardiac arrest, and sesis. The third included all atients who required MV due to altered mental status (Glasgow Coma Score of 8) for any reason. The last category consisted of all neuromuscular rimary causes of resiratory failure, acute or chronic. Variables robably related to PMV were recorded daily: 1) resence of shock, i.e., sesis and setic shock, based on Consensus Conference criteria (5, 6), and cardiogenic and hyovolemic shock, 2) utilization of invasive devices and associated comlications, 3) organ dysfunctions (5, 6), 4) resiratory events, i.e., barotrauma considered to be related to ventilatory management, not including neumothorax secondary to chest trauma or to insertion of central lines, 5) ARDS, based on the American-Euroean Consensus Conference of ARDS criteria (7, 8), 6) ventilator-associated neumonia (defined by the Centers for Disease Control and Prevention diagnostic criteria, namely, temerature of 38.5 C or 36 C, white blood cell count of 12,000 or 4000 cells/mm 3, urulent bronchial secretions, and alveolar infiltrates), 7) gastrointestinal events, namely, enteral or arenteral feeding and gastrointestinal bleeding when hemodynamic suort or red blood cell transfusion was necessary, 8) transfusions of red blood cells ( 30 ml/kg), lasma, or latelets, 9) neuromuscular events, mainly critically ill atient neuromuscular illness and the related drugs, 10) drugs administered (number of days of utilization and dose). Our main outcome measure was the resence of PMV, as defined by MV suort beyond 21 days, and, as a secondary outcome, mortality. Table 1. Patient characteristics at admission and outcome of mechanical ventilation Characteristic or Outcome Nonrolonged Mechanical Prolonged Mechanical Ventilation n 23 Female, n (%) 93 (40) 7 (30).26 Age (months), median (IQR) 6 (2 16) 8 (3 36).15 At least one chronic comlex condition 67 (29) 10 (43).156 Genetic condition, n (%) 20 (9) 6 (26).02 Weight (kg), median (IQR) 6.7 (4.5 10) 6.9 ( ).76 Paediatric Index of Mortality 2 score 7 ( ) 11 (2 22.6).45 (oints), median (IQR) Reason for mechanical ventilation, n (%) Acute on chronic resiratory failure 15 (6) 5 (22).03 Altered mental status 29 (12) 3 (13).03 Neuromuscular 3 (1).03 Acute resiratory failure 186 (80) 15 (65).03 Postoerative 14 (6) 1 (4) 1 Uer airway obstruction 7 (3) 1 (4).5 Acute viral bronchiolitis 89 (38) 6 (26).3 Community neumonia 28 (12) 2 (9) 1 Setic shock 36 (15) 5 (22).3 Outcome of mechanical ventilation Days of mechanical ventilation, 6 (4 10) 35 (26 90) median (IQR) Mortality, n (%) 49 (21) 10 (43) 0.05 IQR, interquartile range. Statistical Analysis. Data are shown as the mean and standard deviation, median with interquartile range, and roortions, as aroriate. All categorical variables were analyzed with the chi-square test, excet where small sizes required the use of Fisher s exact test. The comarison of continuous variables was made with Student s t test for variables with normal distribution or the Mann-Whitney U test for variables with nonnormal distribution. The Kruskal-Wallis test was used to comare continuous variables among more than two grous. Multivariate logistic regression analysis was used to identify factors indeendently associated with PMV. We included the following variables that were significant in the univariate analysis (.01): central venous catheter, ventilator-associated neumonia, setic shock during PICU stay, doamine, use of noreinehrine, use of multile antibiotics, red blood cell transfusion requirement over 30 ml/kg. RESULTS During the study eriod, 256 atients were included and 23 (9%) who required MV for 21 days were assigned to the PMV grou. The median (interquartile range) duration of MV was 6 (4 10) days for the non-pmv grou and 35 (26 90) days for the PMV grou (.001). Patients requiring PMV had higher mortality than atients in the control grou (43% vs. 21%,.05) and longer PICU stay, 10 (6 14) days vs. 35 (28 64) days,.001. No differences were observed in the resective Paediatric Index of Mortality 2 scores, ages, and gender distributions. Only the rate of genetic diseases was higher in PMV atients (Table 1). There was a higher rate of utilization of invasive devices in the PMV grou, but the only comlication that showed a higher incidence associated with their use in the PMV grou was urinary infection (17% vs. 5%,.05). Although there was a higher incidence of heatic dysfunction in the PMV grou (13% vs. 3%,.05), there was no difference in the incidence of other organ dysfunctions. The rates of ARDS and ventilatorassociated neumonia were significantly higher in the PMV grou (43% and 35%, resectively) than in the non-pmv grou (20% and 9%, resectively,.01). There were higher rates of sesis and setic shock during the eriod of MV for atients in the PMV grou when comared to the non-pmv grou. No differences were observed between the two grous in the rates of other tyes of shock (Table 2). 2

3 Table 2. Device utilization, nosocomial infections, and organ dysfunction Device, Infection, or Organ Dysfunction There were more atients in the PMV grou who required doamine or noreinehrine and treatment with multile antibiotics (defined as two or more schemes of antibiotics). No differences were observed between the two grous in the rates of insulin utilization, enteral nutrition, and gastrointestinal bleeding (Table 3). There was a difference between the PMV grou and the non-pmv grou in the ercentage of atients who received a transfusion with red blood cells beyond 30 ml/kg (52% vs. 14%, resectively,.001). Although the roortions of critical illness olyneuroathy/ myoathy were similar in both grous, there was a higher utilization of neuromuscular blockers and corticosteroids in the PMV grou (48% and 84%, resectively) vs. the non-pmv grou (28% and 47%,.05) (Table 3). There was no difference between the two grous with Nonrolonged Mechanical Prolonged Mechanical Ventilation, n 23 Central venous catheter 158 (68; 62 74) 22 (96; ).01 Catheter-associated infection 14 (6; 4 10) 3 (13; 3 32).20 Dee venous thrombosis 5 (2; 1 5) Urinary catheter 204 (88; 83 91) 23 (100; ).07 Urinary infection 12 (5; 3 9) 4 (17; 6 37).05 Ventilator-associated neumonia 20 (9; 6 13) 8 (35; 18 56).01 Barotrauma 10 (4; 2 8) 1 (4; 0 20).65 Acute resiratory distress syndrome 46 (20; 15 25) 10 (44; 25 64).01 Sesis 27 (12; 8 16) 11 (48; 28 68).01 Setic shock 80 (34; 28 41) 20 (87; 68 97).01 Glasgow Coma Score 7 (3; 1 6) 1 (4; 0 20).53 Coaguloathy 3 (1; 0 4) Renal dysfunction 4 (2; 1 4) Heatic dysfunction 7 (3; 1 6) 3 (13; 3 32).05 Hemodynamic dysfunction 110 (47; 41 54) 14 (61; 40 79).21 Values are exressed as n (%; 95% confidence interval). Table 3. Utilization of drugs, antibiotics, arenteral nutrition, and blood roducts during the study eriod Drug, Antibiotic, Nutrition, or Blood Product Nonrolonged Mechanical Prolonged Mechanical Ventilation, n 23 Doamine 98 (42; 36 49) 18 (78; 58 92).001 Dobutamine 33 (14; 10 19) 3 (13; 3 32).59 Einehrine 21 (9; 6 13) 1 (4; 0 20).388 Noreinehrine 34 (15; 11 20) 14 (61; 40 79).001 Milrinone 14 (6; 4 10) 3 (13; 3 32).19 Insulin 14 (6; 4 10) 1 (4; 0 20).60 Multile antibiotics 47 (20; 15 26) 19 (83; 63 94).001 Red blood cell transfusion of 33 (14; 10 19) 12 (52; 32 72) ml/kg Plasma 11 (5; 3 8) 3 (13; 3 32).12 Platelets 16 (7; 4 11) 4 (17; 6 37).09 Parenteral nutrition 10 (4; 2 8) 3 (13; 3 32).10 Neuromuscular blockers 65 (28; 22 34) 11 (48; 28 68).05 Corticosteroids 109 (47; 40 53) 17 (74; 53 89).05 Values are exressed as n (%; 95% confidence interval). resect to sedative and analgesic drug utilization (median and standard deviation): midazolam, 4.8 (7.2) vs. 5.6 (3.5),.08; hentanyl, 64.4 (82.4) vs (64.7),.26 (non-pmv vs. PMV grou). As exected, the roortion of extubation failure and the need for reintubation was higher in the PMV grou: 52% (95% confidence interval [CI] 32% to 72%) vs. 23% (95% CI 18% to 29%),.01. However, the rates of nonlanned extubation were similar in both grous. In all those atients who were on MV for 21 days, four factors were found to be indeendently associated with PMV: setic shock (odds ratio 6.1, 95% CI ,.05), administration of noreinehrine (odds ratio 5.0, 95% CI ,.01), ventilator-associated neumonia (odds ratio 4.8, 95% CI ,.01), and the administration of multile antibiotics (odds ratio 11.0; 95% CI ,.01). DISCUSSION In this study atients who required PMV made u about 10% of all those atients who received MV for 12 hrs. Not only did these atients stay longer in the PICU and require rolonged MV with a greater consumtion of healthcare resources, they also resented a higher rate of life-threatening comlications with increased mortality. The little information on PMV in ediatrics ublished to date is often heterogeneous. The rate of PMV in our study (9%) was similar to rates in studies on adult atients (1, 2, 9). Traiber et al (3) have recently ublished a retrosective study in unselected critically ill ediatric atients showing an incidence of PMV of 2.5% with a mortality of 48%. Although the mortality rate in their study was similar to ours, we found it somewhat difficult to comare both studies because 1) the incidence of PMV was calculated on the basis of all PICU admissions, and not only mechanically ventilated atients, 2) the severity uon admission to the PICU was not reorted, and 3) the oulation was younger than ours. An association between severity scores and PMV has been shown in adults but not in children (10, 11). In this study, the Paediatric Index of Mortality 2 (12) was used because it had shown adequate discrimination for death and survival in an Argentine PICU (13). We found that it was a oor indicator of PMV atients. These did not show greater severity of illness at admission than those who received ventilatory suort for a shorter eriod. They did show, however, differences in interventions during the eriod of MV. The observed mortality rate was higher than that indicated by the Paediatric Index of Mortality 2 score (standardized mortality ratio 1.5), but since the oulation here consisted only of those atients who were mechanically ventilated, this ratio may well be misleading. While no differences were identified at the time of admission, the course of illness for atients in the PMV grou was markedly different from that of atients in the other grou. The incidences of ARDS, ventilator-associated neumonia, and setic shock, all of which are risk factors for PMV in adult atients, were much higher in the first grou than in the second. 3

4 Extubation failure has been shown to range from 5% to 18% (14 17). In our study the overall rate of extubation failure was 25% (95% CI 20% to 31%). Although this seems to be higher, the 95% CI shows a certain overla with recently ublished studies. In one of these an extubation failure rate of 15% (95% CI 6% to 29%) was shown in children with acute viral bronchiolitis (18), and in another, one which described the eidemiologic features of critically ill children with influenza A H1N1 in 2009 (19), the incidence was 24% (95% CI 16% to 34%). At any rate, the high reintubation rate could be related to events during MV, and our findings suggest that the higher rate of utilization of neuromuscular blocking agents and corticosteroids might be one of these events. This association between extubation failure and some muscle-toxic drugs (neuromuscular blocking agents, corticosteroids) is difficult to verify due to the very low incidence in our oulation of critical illness olyneuroathy. While critical illness olyneuroathy has never been recognized as a cause of PMV nor of the failure to wean atients from MV in ediatric oulations (20), it is a wellknown risk factor for PMV in the literature on adult atients (21 25). We did not observe an association between PMV and critical illness olyneuroathy, but this may have been due to misdiagnosis or the small samle size. One of the roblems we faced early in this study was how to define PMV. Previous study oulations varied greatly, from children who had received cardiovascular surgery (26 28) to those who had undergone correction for scoliosis (29) and/or who had received a liver translant (30), etc. In these scenarios, all of which were surgical, PMV was oorly characterized as beyond regular eriod. Information on clinical cases in a PICU is scarce. We were thus obliged to choose one of the following: 1) take the median of duration of MV lus two standard deviations to define the cut oint for PMV, 2) establish the 95th ercentile of duration of MV and consider as PMV every atient ventilated beyond that limit, 3) set a cut oint at 21 days as in adult oulations, 4) set a cut oint at 14 days, consistent with the shorter duration and eidemiology of MV in ediatric oulations (17), or 5) the moment when a tracheostomy is erformed. We chose to use a cut oint of 21 days, as did Traiber et al (3), for the sake of simlicity and because its universality would make it ossible to aly our study results to other oulations. We also erformed an analysis at day 14 (data not shown), which showed a surrisingly similar rofile. Our study has some limitations: The first is a short enrollment eriod. In the context of a changing eidemiology, diagnosis at admission can vary in accordance with the season in which a study is carried out. The second is a aucity of data collection beyond day 21. Although this might have made it ossible to better describe the outcomes of ediatric PMV atients, it was not within the scoe of the study s objective. The third is the rates of extubation failure and mortality, which were higher than those in the bibliograhy, may reflect different inclusion criteria and a selection bias. These limitations, however, should be weighed against the fact that this is a multicentric study carried out in third-level PICUs. In conclusion, this study resents a descrition of a subgrou of critically ill ediatric atients who were mechanically ventilated for a rolonged eriod of time. The mortality rate of these atients was double that of those whose MV was not rolonged. In addition, the survivors had significantly longer PICU stays. At admission, this subgrou could only be differentiated from other atients on the basis of a higher incidence of genetic diseases, but during the course of MV, they suffered more eisodes of ARDS, infections, and setic shock. The identification of risk factors resented here could hel in the imlementation of secific strategies to diminish morbidity and mortality for these atients. REFERENCES 1. Estenssoro E, Reina R, Canales HS, et al: The distinct clinical rofile of chronically critically ill atients: A cohort study. Crit Care 2006; 10:R89 2. MacIntyre NR, Estein SK, Carson S, et al: Management of atients requiring rolonged mechanical ventilation: Reort of a NAMDRC consensus conference. Chest 2005; 128: Traiber C, Piva JP, Fritsher CC, et al: Profile and consequences of children requiring rolonged mechanical ventilation in three Brazilian ediatric intensive care units. Pediatr Crit Care Med 2009; 10: Feudtner C, Christakis DA, Connell FA: Pediatric deaths attributable to comlex chronic conditions: A oulation-based study of Washington State, Pediatrics 2000; 106: Goldstein B, Giroir B, Randolh A, et al: International ediatric sesis consensus conference: Definitions for sesis and organ dysfunction in ediatrics. Pediatr Crit Care Med 2005; 6: Brilli RJ, Goldstein B: Pediatric sesis definitions: Past, resent, and future. Pediatr Crit Care Med 2005; 6:S6 S8 7. Bernard GR, Artigas A, Brigham KL, et al: The American-Euroean Consensus Conference on ARDS. Definitions, mechanisms, relevant outcomes, and clinical trial coordination. Am J Resir Crit Care Med 1994; 149: Artigas A, Bernard GR, Carlet J, et al: The American-Euroean Consensus Conference on ARDS, art 2: Ventilatory, harmacologic, suortive theray, study design strategies, and issues related to recovery and remodeling. Acute resiratory distress syndrome. Am J Resir Crit Care Med 1998; 157: Frutos-Vivar F, Esteban A, Aezteguía C, et al: Outcome of mechanically ventilated atients who require a tracheostomy. Crit Care Med 2005; 33: Seneff MG, Zimmerman JE, Knaus WA, et al: Predicting the duration of mechanical ventilation: the imortance of disease and atient characteristics. Chest 1996; 110: Saijaszko MJ, Brant R, Sandham D, et al: Nonresiratory redictor of mechanical ventilation deendency in intensive care unit atients. Crit Care Med 1996; 24: Slater A, Shann F, Pearson G, et al: PIM2: A revised version of the Paediatric Index of Mortality. Intensive Care Med 2003; 29: Eulmesekian PG, Pérez A, Minces PG, et al: Validation of ediatric index of mortality 2 (PIM2) in a single ediatric intensive care unit of Argentina. Pediatr Crit Care Med 2007; 8: Kurachek SC, Newth CJ, Quasney MW, et al: Extubation failure in ediatric intensive care: A multile-center study of risk factors and outcomes. Crit Care Med 2003; 31: Farias JA, Alía I, Retta A, et al: An evaluation of extubation failure redictors in mechanically ventilated infants and children. Intensive Care Med 2002; 28: Baisch SD, Wheeler WB, Kurachek SC, et al: Extubation failure in ediatric intensive care incidence and outcomes. Pediatr Crit Care Med 2005; 6: Randolh AG, Wyij D, Venkataraman ST, et al: Effect of mechanical ventilator weaning rotocols on resiratory outcomes in infants and children: A randomized controlled trial. JAMA 2002; 288: Johnston C, de Carvalho WB, Piva J, et al: Risk factors for extubation failure in infants with severe acute bronchiolitis. Resir Care 2010; 55: Farias JA, Fernández A, Monteverde E, et al: Critically ill infants and children with influ- 4

5 enza A (H1N1) in ediatric intensive care units in Argentina. Intensive Care Med 2010; 36: Williams S, Horrocks IA, Ouvrier RA, et al: Critical illness olyneuroathy and myoathy in ediatric intensive care: A review. Pediatr Crit Care Med 2007; 8: Latronico N, Fenzi F, Recuero D, et al: Critical illness myoathy and neuroathy. Lancet 1996; 347: Bolton CF: Sesis and the systemic inflammatory resonse syndrome: Neuromuscular manifestations. Crit Care Med 1996; 24: van Mook WN, Hulsewé-Evers RP: Critical illness olyneuroathy. Curr Oin Crit Care 2002; 8: Fletcher SN, Kennedy DD, Ghosh IR, et al: Persistent neuromuscular and neurohysiologic abnormalities in long-term survivors of rolonged critical illness. Crit Care Med 2003; 31: Herridge MS, Cheung AM, Tansey CM, et al: One-year outcomes in survivors of the acute resiratory distress syndrome. N Engl J Med 2003; 348: Székely A, Sái E, Király L, et al: Intraoerative and ostoerative risk factors for rolonged mechanical ventilation after ediatric cardiac surgery. Paediatr Anaesth 2006; 16: García-Montes JA, Calderón-Colmenero J, Casanova M, et al: Risk factors for rolonged mechanical ventilation after surgical reair of congenital heart disease. Arch Cardiol Mex 2005; 75: Lóez-Herce Cid J, Leyton Avilés P, Urbano Villaescusa J, et al: Risk factors for rolonged mechanical ventilation after cardiac surgery in children. Med Intensiva 2008; 32: Yuan N, Skaggs DL, Dorey F, et al: Preoerative redictors of rolonged ostoerative mechanical ventilation in children following scoliosis reair. Pediatr Pulmonol 2005; 40: Glanemann M, Langrehr JM, Müller AR, et al: Incidence and risk factors of rolonged mechanical ventilation and causes of reintubation after liver translantation. Translant Proc 1998; 30: