After receiving approval from the Ethics Committee of Jagiellonian University, Cracow, Poland (approval number: 122.6120.142.2016; 23 June 2016), this prospective observational study was conducted in a single university hospital. All study participants gave written, informed consent to participate in the study before enrolment.
Patient participation was voluntary. The study population was composed of female patients scheduled for elective, gynecological procedures using general anesthesia and positive pressure ventilation. Patients chosen for this study were required to be 18 years of age or older and able to provide informed, written consent. Patients were excluded from this study if receiving pregnancy-related surgery, if the research personnel was unable to obtain all LUS assessments (i.e., the initial assessment before induction of anesthesia or any of the following within the study protocol time frames) or if patients were experiencing any respiratory or circulatory complications at the time of the study. Parameters of interest were assessed using standard perioperative monitoring devices (the Datex Ohmeda S/5 Aespire Patient Monitor, GE Healthcare Helsinki, Finland).
Data collection
Baseline patient characteristics were collected including age, weight, height, body mass index (BMI), ideal body weight (IBW) according to the Lorenz formula [8], American Society of Anesthesiologists Physical Status (ASA) [9], positive end expiratory pressure (PEEP), peak inspiratory pressure (PIP), perioperative fluids management, fraction of inspired oxygen (FiO2) during induction and recovery from anesthesia, intra-abdominal carbon dioxide pressure during laparoscopy, type of surgery (laparoscopy, laparotomy, hysteroscopy), airway management (tracheal intubation versus supraglottic airway), patient’s use of muscle relaxants, intravenous fluid volume, and duration of anesthesia. The LUS and parameters of interest were recorded before induction of anesthesia, at induction of anesthesia and 30 and 60 minutes after induction according to duration of surgery. The last LUS assessment was conducted in the recovery room within two hours after surgery.
Anesthesia
Since it was observational study no changes were done to the anesthesia and intraoperative ventilator protocols and management used in the institution. Thirty minutes before induction of anesthesia, patients received 2 mg of midazolam intravenously. General anesthesia was induced with propofol 2–3 mg kg − 1, fentanyl 1–2 µg kg − 1 and, in cases of muscle relaxation, rocuronium 0.6 mg kg− 1. Maintenance of anesthesia was performed with oxygen/air mixture (FiO2 0.4) and sevoflurane (0.8–1.2 of age adjusted expiratory minimal alveolar concentration), fresh gas flow 3 l min− 1. Ventilation rate (volume-controlled ventilation, I:E ratio of 1:2) was adjusted to maintain normocapnia with a tidal volume of 6–8 ml kg − 1 IBW and PEEP of 2–5 cm H2O. An infusion of intravenous crystalloid fluid therapy was initiated on each patient before induction of anesthesia. Patients received 0.5-1 mg of atropine and 2 mg of neostigmine intravenously to reverse the neuromuscular block.
Ultrasound protocol
There are various LUS assessment protocols that allow qualitative and quantitative assessment of dynamically changing LUS artefacts [10–12]. However, not all of these protocols are relevant to surgery in the supine position. For this reason, the 8-zone protocol, as defined in the “International evidence based recommendations for point of care lung ultrasound” [13], was chosen for this study. All patients’ lungs were assessed for the presence of lung sliding, A-lines, B-lines, interstitial syndrome (two or more bilateral positive regions [presence of three or more B-lines]), lung consolidation and pleural effusion using the 8-zone protocol [13]. For the initial and last LUS assessment, the patients were in a semi-recumbent position. During anesthesia the patients were positioned in the Trendelenburg position according to their specific surgical needs.
All ultrasound assessments were conducted independently by two certified physicians (PK and AJ) experienced in the administration of LUS assessments. All ultrasound assessments were performed using the SignosRT instrument (Signostic Limited, Clovely Park SA, Australia) with a sector probe with lung pre-set and 10 cm of depth (3.0–5.0 MHz; Signostic Limited, Clovely Park SA, Australia).
Statistical analysis
A power analysis was conducted using data from preliminary study. A sample of 56 patients showed an overall 80% power for detecting the differences with two tailed alpha of 0.05, as represented by the presence of three or more B-lines, 30 minutes after the "before induction" time point. In cases of normally distributed data, continuous data are presented as a mean value with standard deviation. Non-normally distributed data are presented as a median value with interquartile range. Normality was assessed using the Shapiro-Wilk test. Discrete data are presented as frequency and percentage.
The changes in number of fields for A-lines, B-lines and ≥ 3 B-lines were assessed by pairing Wilcoxon tests with Bonferroni corrections over multiple comparisons. Comparisons of two repeated measures of dichotomic variables were made using the McNemar test. Analyses were conducted using R software (ver. 3.5.1; R Development Core Team, Austria, Vienna) [14]. Results with a P-value < 0.05 were considered significant.