Indian Pediatrics 2001; 38: 1154-1159
Acute Respiratory Distress Syndrome: Experience at a Tertiary Care Hospital
Acute respiratory distress syndrome (ARDS) is a severe clinical syndrome of acute lung injury(1). Despite a large number of initiating factors, the pathophysiologic events are similar: pulmonary hypertension, intra-pulmonary shunting with severe hypoxemia and myocardial dysfunction. This syndrome was first described in adults(2) and is now increasingly being recognized in children. Almost all the pediatric reports are from the developed countries(3-9). The causes of ARDS in these series are comparable. There are no reports on the magnitude of problem of this entity and its predisposing causes in the pediatric age group in India. The present study was undertaken to examine the incidence rate of ARDS in our pediatric intensive care unit (PICU), to determine the predisposing causes and the outcome.
Subjects and Methods
The records of the children admitted to the Pediatric Intensive Care Unit (PICU) of our hospital during January 1998 to June 2000 were reviewed. The American European Consensus Conference Committee criteria were used to diagnose ARDS: (a) acute onset, (b) bilateral infiltrates on chest radiography, (c) arterial oxygen tension/ fraction of O2 in the inspired gases, i.e. PaO2/FiO2 ratio less than 200, and (d) absence of clinical evidence of left atrial hypertension(10). The medical records of all children so identified were examined to collect information about the triggering event for ARDS, past medical history and other organ failures/dysfunctions. The details of the ventilatory management were analyzed: the type of ventilator used and the ventilator setting used: FiO2, peak inspiratory pressure (PIP), positive end expiratory pressure (PEEP), tidal volume (ml/kg). The blood gas parameters were analyzed using the equipment - AVL 995 (Hersteller 802, Austria) and ABL 50 (Radiometer, Copenhagen). The FiO2 was measured using Partner Oxygen analyzer OA 2100 (Health Care, Taiwan). The following values were calculated. PaO2/FiO2 ratio and the arterial - alveolar oxygen tension difference (AaDO2). The lung injury score was calculated on the basis of chest radiograph, hypoxemia score (PaO2/FiO2), and the PEEP used(11). Respiratory system compliance was not used for calculation because of non-availability of values in most of the cases. Evidence for pulmonary air leak was recorded.
Statistical calculations were performed using the intercooled version of STATA 6.0 (STATA Corporation, TX). Mean (standard deviation) was calculated for various parameters. The features of the survivors and the non-survivors were compared. Fisher’s exact test was applied for discrete variables while Student’s ‘t’ test was applied for continuous variables to test the significance of differences.
During the two and a half year period 992 children were admitted in our 6-bedded PICU [1998: 524, 1999: 361, and 2000 (till June): 107)]. Our PICU predominantly caters to the medical problems. Two hundred and sixteen (21.8%) children were admitted for a respira-tory illness. The other important disease categories were neurological (13.2%), renal (10.9%), and gastroenterological-hepatic (8.8%). One hundred and twenty two (12.3%) children were admitted with shock. In these two and a half years, 20 (12 girls) children satisfied the diagnostic criteria for ARDS. The incidence rate of ARDS in our PICU was 20.1/ 1,000 admissions. In 1998, 6 children were diagnosed to have ARDS, 8 in 1999, and 6 in 2000 [till June]. The mean (SD) age of these children was 52.6 (57.3) months [range 1.3-156 months].
These children had been acutely ill for a mean duration of 8.8 days. Six children had some underlying illness - 3 were undergoing treatment for acute lymphoblastic leukemia, 1 had chronic liver disease, 1 steroid resistant nephrotic syndrome and 1 toxic epidermal necrolysis.
The largest subgroup was comprised of 13 children with pneumonia (including 1 child with aspiration). Three of the above had staphylococcal pneumonia. Seven children had shock as the predisposing cause; of these six had septic shock and one had dengue shock syndrome. The mean interval between the triggering cause and the diagnosis of ARDS was 2.2 (1.4) days.
All children were ventilated. Pressure-controlled ventilators (Sechrist Infant Ventilator) and volume-controlled ventilators (Evita 2, DRAGER) were used in 10 children each. The mean (SD) duration of ventilation was 6.8 (6.2) days [range 1-21 days]. The maximum PEEP used during the course of ventilation was 7.2 (2.6) cm H2O [range 4-12], while the maximum PIP used was 35 (7.9) cm H2O (range 30-45). All children received 100% oxygen in the beginning. The minimum FiO2 used during ventilation was 71 (27%) [range 30-100%]. In children who were ventilated using a volume-controlled ventilator, the mean tidal volume used was 11.8 (4.1) ml/kg [range 6.9-20 ml/kg]. In five children inverse ratio ventilation was also used to optimize the oxygenation.
The minimum PaO2/FiO2 ratio during the ventilatory support was 60.7 (29.2) [range 19.6-110.1]. The maximum AaDO2 was 577.2 (25.9) mm Hg [range 291.1-631]. The maximum PaO2 achieved was 126 (63.8) mm Hg [range 19.6-321.5 mm Hg]. The maximum PaO2 value during the course of ventilation was 60 (25.9) mm Hg [range 36.9-109]. The mean lung injury score calculated using three parameters(11) was 2.95 (0.43).
During ventilatory support, 3 children developed pneumothorax. All these children had staphylococcal pneumonia. The maxi-mum PEEP used was 4, 5, and 5 cm, respect-ively. The lung injury scores were 2.67, 3, and 2.67, respectively. Of these, two children survived.
Fifteen (75%) children died. The major cause of death was refractory hypoxemia due to ARDS (in 8 children). Septic shock contributed to death in seven children.
The mean duration of ventilatory support in survivors [n = 5] was 11(7) days [range 4-21 days]. These children stayed in hospital for 34 (22) days [range 11-62 days]. We compared the various parameters in the survivors and the non-survivors (Table I). The differences in the age and sex distribution were not statistically significant. All the survivors had pneumonia as the predisposing causes while 53.3% of non-survivors had developed ARDS consequent to shock; this difference did not reach statistical significance. The survivors had statistically significant higher values for maximum and minimum PaO2, minimum PaO2/FiO2. How-ever, the AaDO2 was similar in the two groups, Statistically significant lower PEEP was used in the survivors, while the PIP was similar. In children ventilated using a volume-controlled ventilator, the non-survivors received higher tidal volumes; this difference did not reach significance.
We observed the incidence rate of ARDS in our PICU was 20.1/1000 admissions. The most common underlying cause was found to be pneumonia. The mortality rate was 75%.
In western adult populations, the reported incidence varies from less than 5 per 100,000 to 75 per 100,000 population(12,13). In children the population-based data is not available. The reported rates in PICU vary from 8.5 to 27 cases/1000 admissions(5-9). Our preliminary experience suggests that rate in our PICU was 20.1 per 1000 admissions for a two and a half year period. Our initial experience suggests an increase in incidence over a period of time. The possible reason for observing higher incidence rates could be greater number of admissions with infectious diseases. Because of the limited number of beds and a greater proportion of ventilated children, the tendency may have been to admit more sick children, especially those at risk of developing ARDS. Most of the PICUs in the developed countries cater to both medical and surgical patients while our PICU does not cater to children with surgical problems. This may also contribute to the higher incidence rates. The increase in the rates with time may be due to greater aware-ness and also due to increase in the severity of illness in children admitted to PICU. There is no pediatric data available from developing countries like India. This could be due to poor awareness and poor availability of facilities for blood gas analysis and ventilation.
Figures are mean (SD). Statistical test used: *Fisher’s exact test; for other parameters Student’s ‘t’ test was used. * As defined by Murray et al.(11).
The predisposing causes can be divided into two groups; those causing direct injury to the lungs and those that cause injury indirectly in the setting of a systemic process(1). The common causes associated with direct lung injury are pneumonia and aspiration; near drowining, embolishm and pulmonary contu-sion are less common(4-9). Some authors have excluded children with pneumonia when reporting on ARDS(3); however, it is now well accepted that pneumonia is a predispos-ing factor for ARDS(1,9). We observed that pneumonia was the risk factor in 12 children while aspiration was responsible for ARDS in one child.
Sepsis is the commonest cause of indirect lung injury(1). Severe trauma and acute pancreatitis are less commonly associated causes. Septic shock was the predisposing cause in 6 children while one had dengue shock syndrome. These results are similar to those in other pediatric series(3-9). The number of subjects in these reports varies from seven to sixty.
These children require aggressive ventila-tory support. A PEEP of greater than 5 cm H2O was used in all the patients except when pneumothorax had occurred. There was a tendency to use high tidal volumes if the child did not improve. However, with increasing experience and confidence, we are using higher PEEP with lower tidal volumes. A recent study has conclusively shown that low tidal volume ventilation significantly reduces the mortality in adults with ARDS(14). The current strategy should be to use low tidal volume ventilation and use high PEEP to ensure optimal oxygenation. The target PaCO2 could be higher, about 50 mm Hg while the target PaO2 should be about 60- 80 mm Hg.
The high mortality rate (75%) observed by us is comparable to that reported earlier (28.5-75%) (3-9). The majority of our patients had severe disease as reflected in high lung injury scores and low PaO2/FiO2 ratios. We compared the characteristics of the survivors and non-survivors. The non-survivors appeared to have a more severe disease, as evidenced by lower values of PaO2, and lower PaO2/FiO2 ratios. Higher PEEP and higher tidal volumes were used in non-survivors; this may reflect the attempts to improve oxygena-tion in these children with more severe disease. In addition, higher tidal volumes may have contributed to the lung injury(15,16). Thirty eight per cent of the children with pneumonia as the predisposing cause survived while none of the seven children with shock survived. This may be due to presence of systemic process, which can affect all the organ systems, in the latter cate-gory. However, because of the low number of patients definite conclusions cannot be drawn. Available literature suggests that presence of sepsis, non-pulmonary organ dysfunction and chronic liver disease increase the risk of mortality significantly(17,18). A study in the pediatric age group found significant differences between the survivors and non-survivors in intrapulmonary venous admixture, alveolar-arterial oxygen tension difference, oxygenation index, and peak inspiratory pressure(4). Another study found that the non-survivors had higher minute ventilation, maximum PEEP, maximum FiO2, and maximum PIP compared with the survivors(9). The authors also observed that AaDO2 >420 was the best early predictor of death in children with ARDS(9). The lung injury severity scores(11) do not seem to predict mortality.
Our preliminary experience suggests that ARDS in children is not uncommon. Efforts should be made to diagnose this entity.
RL and SKK collected the data and drafted the manuscript. RMP analyzed