A prospective cohort study was conducted, and children with moderate to severe ARDS admitted to pediatric intensive care unit (PICU) were enrolled from November 2016 to October 2019 at Shanghai Children's Hospital. Moderate to severe ARDS was defined according to the Berlin definition of ARDS (15). The exclusion criteria included: 1) patients who were in PICU less than 72 hours; 2) patients with lack of appropriate acoustic window; 3) patients with pneumothorax; 4) patients with hypoxemia secondary to cardiac disease congenital cardiovascular disease or chronic cardiopulmonary disease. According to whether CRRT were used during PICU stay, patients were divided into CRRT group and Non-CRRT group.
All the patients were received the mechanical ventilation based on lung protective-ventilation strategy or/and prone positioning, neuromuscular blockade (NMB), conservative fluid management (17). The total fluid volume was generally 60-70ml/kg.d or 1200-1500ml/m2.d, blood transfusion if hemoglobin level down to 7.0g/dL during PICU stay (16). Patients received diuretics according to daily fluid balance when patients didn’t receive CRRT.
The study protocol was approved by the local ethics committee of Children's Hospital affiliated to Shanghai Jiao Tong university (Approval number:2016R007-E03). The informed consent was signed by the patients’ parents or relatives.
Lung ultrasound score
LUS was performed using a 13-6 MHz transducer (M-Turbo Ultrasound System, Mini-Dock-M Series, SonoSite). According to previous studies (18-21), patients were examined in supine, lateral, and prone positions applying the probe perpendicularly to the chest wall surface in order to get the longitudinal scan. Each hemithorax is divided into three regions by sternum, anterior and posterior axillary lines, and each region is divided into upper and lower halves. Each region should be correctly identified the pleura lines and A line by the linear probe in longitudinal scan.
Twelve areas are identified in turn and each region is assigned scores from 0 to 3. The LUS score is the sum of twelve areas, and the final LUS ranges from 0 to 36. In the present study, the definition of LUS score (22) and the representative images for different scores were shown (Table 1). The pathophysiological changes were described with different ultrasonic signs (23). All of the images and clips were collected and evaluated by two PICU expert physicians independently who had been trained and could complete lung ultrasound skillfully. All of the images, physical characteristics, baseline data and treatment of patients were all anonymizated when they were evaluated by these operators.
CRRT and mechanical ventilation
The CRRT mode was continuous veno-venous hemofiltration (CVVH) using Prismed or Prismaflex M60/100 membrane hemofilter equipped with an AN69 (Gambro Renal Products, Meyzieu, France) in a multifiltrate continuous renal replacement therapy machine (Gambro or Gambro prismaflex, Gambro Lundia Monitor Division, Lund, Sweden). The indications, performance, and management for CRRT were described as our previous study (14).
The modality of mechanical ventilation was intermittent mandatory ventilation (IMV) with PEEP levels 8-15cmH2O and positive inspiration pressure (PIP) based on target tidal volume (Vt) of 4-8ml/kg (16). Parameters were aligned with lung protective ventilation strategy when patients met the diagnosis of moderate to severe pediatric ARDS.
Demographic data such as age, sex, and body mass index (BMI), the pediatric risk mortality III (PRISM III) score(24) and co-morbidity were collected on PICU admission. Clinical parameters including fractional concentration of oxygen in inspired gas (FiO2), PaO2/FiO2, PaCO2, OI, dynamic lung compliance (Cdyn) which was continuously displayed using ventilators (MAQUET company, Servo-i serious) (25, 26) and daily fluid balance information, hospital mortality were collected. LUS score was determined within 2 hours after moderate to severe ARDS diagnosed as the value of 1st, then measured every morning in following three days as the values of 2nd, 3rd, and 4th. The schematic diagram of LUS score determination was shown in Figure 4. In addition, duration of mechanical ventilation, duration of CRRT, length of PICU or hospital stay was recorded.
The data were performed with SPSS 17.0 statistics (SPSS Inc, Chicago). The characteristics of the patients were reported as percentages for categorical variables and compared the differences between groups by chi-square test. The continuous data with abnormal distribution were expressed as median (interquartile range, IQR) and compared using the Mann-Whitney U test. The correlation between LUS score and PaO2/FiO2, Cdyn, PaCO2, OI and the correlation between the change in daily fluid balance volume and the change in LUS score during CRRT were all performed using Spearman correlation analysis. Friedman test was used to compare mean of more than 2 sets of data. P value < 0.05 was considered to be statistically significant.