The study was conducted at Peking University People’s Hospital in Beijing, China, and Strengthening the Reporting of Observational studies in Epidemiology guidelines were followed. The study was approved by the Institutional Review Board of Peking University People’s Hospital, Beijing, China, and written informed consent were obtained from all subjects participating in the trial. The trial was registered prior to patient enrollment, registration number for clinical trials being ChiCTR1800014508 (www.chictr.org.cn, Jan 17, 2018; Hong Zhao, M.D.). In this observational single-blind study, patients aged between 18 and 75 years who were diagnosed as lung nodules and scheduled for VATS under general anesthesia were included. Patients were excluded if they were allergic to NSAIDs and had a history of asthma, peptic ulcer disease, inflammatory bowel disease, or renal deficiency.

Patients were divided into two groups, the erector spinae plane block group (ESPB group, who received preoperative single shot of bi-level paravertebral block at T4 and T6 levels) and the control group (who received intercostal nerve blocks through T5 to T7 intercostal spaces along mid-axillary line after surgery). Ultrasound-guided ESPB was performed in the preparation room. Patients who did not receive ESPB due to unavailability of preparation room received intercostal nerve blocks performed by the surgeons. The dermatome block level in two groups was tested after extubation in the postoperative care unit by a resident who was unaware of patients’ grouping. Continuous infusion of flurbiprofen (8 mg/h) and intravenous oxycodone rescue (1 mg/bolus, lockout time 10 min) were provided as postoperative analgesics.

Cognitive function was measured for all patients on the day before surgery and 48 h after surgery (on postoperative day 3) using brief Cogstate computerized battery to assess psychomotor speed, visual attention, visual memory, and working memory. Cogstate computerized battery was installed in a laptop and provided to the patients by a resident who was unknown of patients grouping.

Demographic, anesthetic, surgical data, the occurrence of postoperative complications, and the postoperative recovery of patients were collected. The changes between pre- and postoperative cognitive functions and the incidence of delayed neurocognitive recovery 48 h after surgery were analyzed and compared.

Ultrasound-guided ESPB was performed in the lateral position using longitudinal, in-plane technique under strict aseptic precautions in the preparation room. The tip of the T4 transverse process was identified using a high frequency ultrasound transducer by recognizing the fourth rib counting down from C7. The parasagittal view revealed a subcutaneous tissue, trapezius, rhomboid, and erector spinae muscle layers superficial to the transverse processes. With T4 transverse process placed in the middle of the image, a 22-gauge block needle was inserted in-plane and directed to the middle on T4 transverse process. The correct location of the needle tip in the fascial plane deep to the erector spinae muscle was confirmed by injecting 0.5 to 1 mL of saline to visualize lifting of the erector spinae muscle off the transverse process without distending the muscle. Fifteen milliliters of 0.33% ropivacaine was then injected into the erector spinae plane. Another 15 mL of 0.33% ropivacaine was injected into erector spinae plane by identifying T6 transverse process (Fig. 1).

Ultrasound guided erector spinae plane block. A The block needle was advanced from a caudal to cranial direction. B Ultrasonographic image of the T4 transverse process in the middle of the image. C The block needle was advanced in-plane until the needle tip contacted the T4 transverse process (TP). D Spread of local anesthetic was observed in the plane between the erector spinae muscle and the transverse process, and erector spinae muscle lifted during the block. T4TP, transverse process

Patients in the control group received intercostal nerve blocks through T5 to T7 intercostal spaces along mid-axillary line after surgical closure performed by surgeons. One percent ropivacaine 10 mL was injected at the lower border of the fifth (3 mL local anesthetics), sixth (3 mL), and seventh rib (4 mL) along the mid-axillary line.

General anesthesia was induced with etomidate (0.3 mg/kg), sufentanil (0.3 μg/kg), and rocuronium (0.6 mg/kg), and an appropriate size of double-lumen tube was inserted. Anesthesia was maintained with continuous infusion of propofol (4–6 mg/kg·h), remifentanil (0.1–0.2μg/kg·min), and dexmedetomidine 0.4μg/kg·h to accomplish bispectral index ranging from 40 to 60, heart rate and invasive blood pressure varying from baseline ± 20%. Additional doses of sufentanil were left to the discretion of the anesthesiologist in charge to achieve adequate analgesia. Dexmedetomidine was discontinued and ondansetron 5 mg was given intravenously 30 min before the end of surgery.

A loading dose of flurbiprofen (1.5 mg/kg) was given intravenously at the end of surgery and a continuous infusion of flurbiprofen (4 mg/mL) was given at 2 mL/h with a total volume of 100 mL through a single-use infusion device (KSH®, KSH Medical Technology, Beijing, China). Postoperative analgesic regimen also comprised of patient-controlled intravenous oxycodone rescue (CADD®, Smiths Medical, Ashford, UK), set as bolus of 1 mg/2.5 mL (total volume as 40 mg/100 mL), with a lockout time of 10 min.

The total amount of intraoperative sufentanil and remifentanil were recorded. The visual rating scale (VAS) scores of pain and analgesic consumption through the patient-controlled intravenous analgesia (PCIA) pump were followed up at 24 h and 48 h after the operation. Nausea, vomiting, dizziness, and other complications were recorded.

Cognitive function was measured for all patients on the day before surgery and 48 h after surgery using brief Cogstate computerized battery to assess psychomotor speed, visual attention, visual memory, and working memory. Four different tasks provided by CCB including “detection task (DET),” “identification task (IDN),” “one card learning task (OCL),” and “one-back task (OBK)” were used to assess psychomotor speed, visual attention, visual memory, and working memory respectively. The changes between pre- and postoperative cognitive functions and the presence of delayed recovery of cognitive function after surgery were analyzed. Delayed neurocognitive recovery or cognitive decline was defined as 1.96 SD decrease in 2 tests (Evered and Silbert 2018).

The study was initially powered as a prospective cohort study. The sample size was calculated form the results of a pilot study of 10 patients who received general anesthesia with intercostal nerve blocks after surgery. Reaction time on the detection task increased from 2.65 log10 ms (SD = 0.11) ms before surgery to 2.85 log10 ms (SD = 0.15) after surgery. We assumed that preoperative ESPB would result in a smaller increase in reaction time on the detection task (approximately 50%) to a detection task speed of 2.75 log10 ms after surgery. Therefore, 30 patients per group were required (α = 0.05 two tailed, β = 0.2).

Statistical analysis was performed using the SPSS 24.0 statistical software package (SPSS Inc., Chicago, IL, USA). Continuous variables are expressed as mean (SD) or medians with interquartile range and categorical variables as numbers and percentages. Between-group differences were evaluated using the independent t test or Mann-Whitney U test for continuous variables and the chi-square test or Fisher exact test for categorical variables, as appropriate.