Acute Hypoxemic Respiratory Failure In Pediatric Intensive Care Units: A Multicenter Clinical Epidemiologic Study

BACKGROUNDAcute hypoxemic respiratory failure (AHRF) causes significant morbidity and mortality in pediatric intensive care unit (PICU). Persistent hypoxemia may induce multiple organ system failure and requires aggressive respiratory therapy as life support. AHRF may be associated with development of acute lung injury (ALI) and acute respiratory distress syndrome (ARDS).In pediatric AHRF, clinical management and outcome vary in different countries depending on both service standard and socioeconomic conditions. In developed countries, recent studies of pediatric AHRF and ALI reported survival rates of 70%-73% and data from ALI/ARDS clinical trials even revealed survival rates as high as 89%. However, survival rates from those interventional clinical trials couldn't reflect the real condition, because those studies just targeted on specific populations and excluded a part of patients with high risks of death. Recently reported epidemiological studies revealed the mortality of ARDS remained as high as 60%. Data from clinical epidemiologic study should more closely reflect real practice level in the management of pediatric AHRF, and this information is scant in developing countries.Most experience of management of pediatric AHRF and ALI were based on neonatal and adult studies. Unfortunately, most randomized trials were inconclusive except a single-center demonstrated a 33% absolute mortality reduction in patients treated with a "lung protective ventilation strategy" including reduced tidal volumes, pressure-cycled mode and a positive end expiratory pressure (PEEP) adjusted by pressure-volume curve analysis. One year later, a large multicenter clinical trial (American ARDS-Net trial) reveal that ventilation with tidal volume at 6 ml/kg under volume-cycled model and a plateau pressure≤30 cm H2O was associated with an absolute 8.8% decrease in mortality compared with conventional ventilation strategy (12 ml/kg). Other interventions such as corticosteroid therapy, selected PEEP, prone position, fluid management, inhaled nitric oxide, although they could improve lung function and oxygenation, this physiologic improvement could not finally transferred into improvement of outcome. It was speculated that the lack of thorough understanding of this critical condition should contribute to the possible unreasonable designed clinical trials which couldn't resulted in valuable evidences. Therefore, it should be critically important to re-understand or gain deeper understanding of the clinical epidemiologic characteristics of AHRF.Although there were a large numbers of AHRF clinical studies, most of them focus on ALI/ARDS in adults, few of them targeted pediatric patients and none from developing countries. In past decades, with economic booming in China, PICU service has been improved steadily. However, advanced respiratory care as a part of standardized care for AHRF remains incomplete in terms of availability and experience of mechanical ventilation, adjunctive therapy, follow-up, etc. A well understanding of epidemiologic characteristics of pediatric AHRF should enable a better assessment of the feasibility and more reasonable design of clinical trials in this patient and disease entity. In 2000-2005, we conducted 3 multicenter ARDS surveys consecutively. The first one conducted in 15 adult intensive care units (ICU) in Shanghai from March 2001 through February 2002. The incidence of ARDS was 2%; in-hospital and 90 day mortalities were 68.5% and 70.4%. Multiple organ system dysfunction was the most common cause of death (60%). The other 2 studies were conducted in 4 PICUs in Shanghai and 25 PICUs from national-wide, respectively. ARDS accounted for about 1.4% of the total critically ill patients in PICU and associated with mortalities of 61%-71%. Pneumonia (56%) and sepsis (28%) were the most common underlying disease. Compared to pneumonia, sepsis seemed more likely to develop ARDS. In PICU, the mortality of pneumonia patients who developed ARDS was extremely higher than that in pneumonia patients without ARDS, so did in sepsis patients. Most ventilated pediatric ARDS patients were on pressure-cycled model instead of volume-cycled model and low tidal volume lung protective strategy.After those 3 studies, we have had an outline picture of ARDS in adult ICU and PICU. But there are still some concerns unclear:the incidence and mortality of AHRF in PICU; the relationship between AHRF and ARDS; whether the mortality of ARDS changed; how is the respiratory support in PICU now; how are the compliance and consistency of the participants in executing study protocol after last collaborative study.In order to determine those problems quoted above, the present study set up an AHRF collaborative study group based on the previous ARDS collaborative study group and has included the same study protocol as in the previous ARDS. So as we could have a longitudinal comparison with regard to the improvement in the settings and practice of the PICU where most investigators are involved. The aim of the present study was to achieve a longitudinal view in the management of AHRF, ALI and ARDS in PICUs, and understand clinical characteristics of these severe respiratory diseases with respect to predictive values of clinical management and outcome; provide valuable information for further interventional study and randomized clinical trial.OBJECTIVES1. To determine the incidence, causes, mortality and its risk factors, and costs of AHRF and ARDS in infants and children in PICUs of major cities in China.2. To evaluate the respiratory support, fluid management used in AHRF treatment and to investigate predictors of AHRF outcome.3. To determine the incidence and outcome of ALI/ARDS.4. To explore the difference in AHRF patient enrollment and therapy among the collaborative PICUs.METHODSThe study was designed as descriptively prospective clinical epidemiologic study, conducted from Dec.2005 through Nov.2006 for 12 consecutive months. A pediatric AHRF collaborative study group included 26 PICUs was set up based on the previous pediatric ARDS collaborative study group. The 26 PICUs were divided as from economically developed region (13) and underdeveloped region (13). All PICU admissions were screened according to Chinese pediatric critical ill score and American guidelines for admission and discharge policies for PICU, all critically ill patients aged from 29 days to 15 years were eligible to study entry. AHRF was defined as:PaO2≤50 mmHg or PaO2/FiO2≤250 mmHg for≥6 h; or required mechanical ventilation≥6 h (necessitating a FiO2>30% and PEEP>2 cmH2O to maintain PaO2>60 mmHg or SpO2>90%); or PaO2/FiO2 remains≤250 mmHg after 6 h ventilation; ALI/ARDS diagnosed according to the 1994 America-European Consensus Conference definitions (AECC).Data were recorded every 6 hours in first 3 day and than daily, included demographic, physiologic, intervention, medication, outcome and cost. Children's hospital of Fudan University was pointed as coordinative center and in charge of collection of data and communication among different PICUs. All data were analyzed with statistic software SPSS 15.0.MEASUREMENTS AND RESULTSDuring the 12-month study period,16 442 patients were admitted to those 26 PICUs,11 521 identified as critically ill patient. A total of 461 met AHRF criteria, 348 of them fulfilled AECC criteria for ALI and 306 fulfilled ARDS criteria, accounted for 4%(461/11 521),3%(348/11 521) and 2.7%(306/11 521) of total PICU critically ill patients, respectively. ARDS accounted for 66.4%(306/461) of AHRF patients and 88%(306/348) of ALI patients.Median age of AHRF patients was 11 months,50.5%(233/461) were less than 1 year old; patients aged 0-5 years accounted for 83%(381/461) of all AHRF patients. Out of 461 patients,452 (98%) were Han and 324 (70.3%) were male.Most common underlying diseases of AHRF were pneumonia (75.1%) and sepsis (14.7%), followed by near drowning (2.5%), lung contusion (1.4%) and bypass (1.4%). In AHRF patients<1 year old,87% originate from pneumonia, just 10% from sepsis; as age increased, the percentage of pneumonia decreased and sepsis increased. In patients aged 0-5 years and 6-15 years, the percentages of pneumonia origination were 79%and 53%, respectively; and the percentages for sepsis were 12% and 29.One hundred and ninety-two AHRF patients did not survive to discharge, accounted for 15.5% of total death in PICU (192/1241); resulted in a mortality of 41.6%(192/461), which was great higher than that of critically ill patients in PICU (10.8%, P<0.01). There were 33%(153/461) patients discharged because of medical support withdraw. The overall 90-day mortality for AHRF was 41.8%(193/461), which was similar to in-hospital mortality. Among patients aged 0-5 years,156 died in hospital (41%,156/381); mortality of patients 6-15 years old was 45%(P=0.5). Median survival time of non-survival patients was (22-184),28%(55/192) death occurred within 24 hours after enrollment,56%(107/192) in three days.The mortality of pneumonia patients who developed AHRF was extremely higher than that of pneumonia patients without AHRF (42.6% vs 3.6%, P<0.001; OR 17.2,95%CI 13.3-22.7), so did in sepsis patients (58.8% vs 16%, P<0.001; OR 7, 95% CI 4.5-12.6). The predominant cause of death was multiple organ system dysfunction (43%,82/192), followed by underlying disease (39%,75/192) and economic reason (16%,30/192). Total PICU cost of 461 patients was 9 383 858 yuan (CNY,10% of cost of all critically ill patients). Median cost of AHRF was 12 750 yuan,2 times the average of critically ill patients, equal to the urban resident's per capita disposable income and 4 times the rural resident's per capita income in China in the same period.There were 71%(373/461) patients received mechanical ventilation, CPAP 1.6% (53/461) and 35 patients without any assistant ventilation (oxygen supplement). Of the 373 ventilated patients,93%(346/373) ventilated on pressure-cycle model and 5.4%(20/373) on volume-cycled model; Only 207 (55.5%) patients with tidal volume record, median exhaled tidal volume was 8.8 ml/kg (8.0-12), median PIP 24 cmH2O (20.4-26.8), median MAP 11.6 cmH2O (9.2-14.2) and PEEP 4.5 cmH2O (3-6).The fluid intake during the first 3 days was 109 ml.kg-1.d-1 (76-175), in first 7 days was 110 ml.kg-1.d-1 (78-165); the fluid balance (fluid in minus out) were 23 ml.kg-1.d-1 (4-51) and 22 ml.kg-1.d-1 (4-49), respectively. The fluid balance of non-survival patients was higher than that of survivor (31.2 ml.kg-1.d-1 vs 23.8 ml.kg-1.d-1, P=0.032). A total of 277 patients received corticosteroid therapy, average using time was 5 days,20%(56/277) of them>7 days. There was no significant difference between patients had had corticosteroid therapy or not in term of mortality (45% vs 36.4%, P=0.063). Only 5 patients received inhaled nitric oxide therapy.Sepsis origination as well as hepatic, renal and central nervous system dysfunction was independently associated with mortality in AHRF patients. PaO2/FiO2 lower than 200 mmHg for 3 days or longer was associated increased risk of mortality.There were 348 patients met the AECC criteria for ALI,306 (88%) of them developed ARDS. The median age of ARDS patients was 13.5 months (4-49.8),62% (190/306) of them<1 year old,80%(245/306) of them were 0-5 years old. The predominant underlying diseases of ARDS were pneumonia (71%) and sepsis (17%). Compared to pneumonia patients, sepsis patients were more likely to develop ARDS. The in-hospital mortality of ALI and ARDS were 42.2% and 44.8%, respectively. Median survival time of Non-survival ARDS patients were 57.6 hours (19-186). Eighty-seven percent ARDS patients received mechanical ventilation, median PEEP was 4.7 cmH20, PIP 24 cmH20, MAP 11 cmH20 and tidal volume 9.8 ml/kg (7.3-10.5). Total PICU cost of ARDS accounted for 7.8% of that of total PICU critically ill patients. Median cost of ARDS was 1.5 times the urban resident's per capita disposable income and 4 times the rural resident's per capita income.The median percentage of critically ill patients in all admissions was 78.5% (46-90) among the 26 PICUs,4 of them with percentages less than 40%. The number of enrolled patients, incidences and AHRF mortalities varied greatly among different PICUs. When 9 PICUs with enrolled patients<11 patients were excluded, the variation of mortality among the remaining 19 PICUs decreased significantly. The mortality of AHRF in hospital from developed region seemed lower than that in hospital from underdeveloped region, but there was no statistically significant difference. There was no difference between hospital from different regions and whether university hospital or not, in term of respiratory support and fluid management.CONCLUSIONS1. AHRF in PICU is associated high mortality and medical burden. AHRF patients are mainly from population aged 0-5 years. Their construction of underlying diseases is different from that in 6-15 years old populations. The period of first 3 days after diagnosis is the key phase to rescue the patients.2. The predominant cause of AHRF death was multiple organ system dysfunction. Sepsis origination as well as hepatic, renal and central nervous system dysfunction is independently associated with AHRF mortality. PaO2/FiO2 lower than 200 mmHg for 3 days or longer was associated increased risk of mortality.3. The diagnosis rate of ARDS in present study is higher than that in our previous study. The mortality has seen distinct decrease, but there is no change in respiratory support.4. The participants' understanding of ARDS has improved after previous study, but their compliance and consistency need further improvement.