Subjects
This study was carried out at the academic 57-bed PICU, First Hospital, Jilin University in China. Seventy consecutive patients (age less than 18 years) were enrolled with the requirement of invasive MV for more than one full day between January 2019 and January 2020. The hospital’s institutional ethics committee consented to the protocol followed in this study (ChiCTR1800020196). The guardians or parents of the enrolled children were informed of the protocol for which, they gave written consent and were also provided with an information sheet.
Each participant fulfilled the established criteria[15] for readiness for weaning (recovery in the primary disease cause, PEEP or positive end-expiratory pressure ≤ 5-10 cm H2O, PaO2/FiO2 > 200, , FiO2 ≤ 50%, and hemodynamic stability when the vasopressors are lacking). The criteria for exclusion were known neuromuscular disease (like myasthenia gravis, Guillain-Barre, or amyotrophic lateral sclerosis), injury to the cervical spinal cord, pneumothorax, guardians or parents not willing to take part in the study.
Study Design
Ultrasound was performed on enrolled subjects to assess the diaphragm during the SBT (spontaneous breathing trial), carried out using support trials using a ventilator (Drager Evita 4) for half an hour at 5 cm H2O PEEP and 8 cm H2O pressure support. Five min post-SBT initiation, ultrasound measurements were taken. Each enrolled patient was either assigned to the DD- or the control (non-DD) group based on the outcome of diaphragmatic echo. DD was described as a DTF (diaphragmatic thickening fraction) of less than 20% at the time of tidal breathing[16].
Evaluation of the diaphragm ultrasound
The ultrasonography of the diaphragm was done by two experienced sonographers using a movable ultrasound apparatus from Mindray (M7 series, China) using a linear probe of 10HMz. The measurement of only the right hemidiaphragm was done because of better feasibility and repeatability of the right hemidiaphragm than the left hemidiaphragm[12]. The head of the bed was at an angle of 30-degrees and each subject was positioned semi-recumbently. Placement of the probe was done in the 8th to 11th intercostal space, between the antero-axillary or mid-axillary line, and perpendicular to the skin in a cranio-caudal direction to view the right hemidiaphragm properly[17]. The image analysis of the diaphragm ultrasound revealed a hypoechoic structure in the middle of two echoic lines (the peritoneal and the pleural membranes)(Fig 1). In imaging of the B-mode, the estimation of Tdi (diaphragm thickness) was done from the internal edges of the pleural line to that of the peritoneal line at ends of both inspiration and expiration. DTF was calculated as (Tdi-inspiration – Tdi-expiration) / Tdi-expiration[18]. In children, Tdi and BW (bodyweight) are correlated significantly [19]. Therefore, the standardization of Tdi was done by BW (DE/BW).
Characteristics of Patients and Clinical Outcomes
For each enrolled patient, the basic demographic were collected, in addition to the data on primary diagnosis, the levels of inflammatory factors at discharge and the period for which these inflammatory factors were elevated, and medications, because of the association of systemic inflammation with muscle atrophy in adult patients who were suffering from critical illness[20]. We observed the clinical outcomes including difficulty or delay in weaning, failure of extubation, duration on MV, duration of PICU stay and mortality.
Analyses of Statistical data
For clinical characteristics, demography, and DD and non-DD patient group outcomes, the comparison of continuous variables was done through the Mann-Whitney U test or Student t-test. Comparison of categorical variables was done through Fisher’s exact test or Chi-squared test. For continuous variables having normal distribution, presentation of data was done as mean ± standard deviation while that for variables with a non-normal distribution was done as median with interquartile range. For the description of categorical variables, n (%) was used. Logistic regression analysis was done to examine the factors significantly associated with DD. The SPSS Statistics, V22.0 from IBM Corp (Armonk, NY) for Windows was utilized for all analyses, and a p-value ≤ 0.05 was deemed significant statistically.