This prospective, randomized, single-blinded trial protocol was approved by the Ethics Committee of the Second Affiliated Hospital of Harbin Medical University on November 10, 2020 (KY2020-222), and it was registered at www.chictr.org.cnon12/01/2021 (registration No.: ChiCTR2100042037). The methodology of this study follows the Helsinki Declaration (revised, 2014). From November 10, 2020, to May 30, 2021, 230 patients were selected for posterior lumbar decompression or fusion surgery. All patients signed written informed consent before enrollment. Inclusion criteria: age ≥ 60 years, BMI ≤ 32 kg/m2, and American Society of Anesthesiologists (ASA) physical status I to III. The exclusion criteria were as follows: (1) schizophrenia, epilepsy or Parkinson’s disease; (2) inability to communicate or have cognitive dysfunction (Screening of mini-Mental State Examination (MMSE)) [7]; (3) history of traumatic brain injury, stroke or neurosurgery; (4) injection site infection; or (5) contraindications to flurbiprofen. Exit criteria included (1) withdrawal or early discharge from the study; (2) self-administered analgesics; (3) cerebrospinal fluid leaking during surgery; or (4) serious cardiac and cerebrovascular events during the trial.

The anesthesiologist in charge of intraoperative management evaluated the patients on the day before the operation and allocated them to the ESPB or control group based on a 1:1 random number (SPSS Statistics 26.0, IBM, USA). Standard general anesthesia was conducted in the two groups. The ESPB group received T12 ESPB after intubation, whereas the control group did not receive the intervention. T12 ESPB was performed by a specialist who was experienced in nerve blocks. Patients, postoperative physicians and postoperative investigators were unaware of the group assignments throughout the trial.

The electrocardiogram (ECG), pulse oxygen saturation, and bispectral index (BIS) were monitored once the patient entered the operating room. Depending on the patient’s condition, invasive or noninvasive arterial blood pressure monitoring was performed. Intravenous (IV) sufentanil (0.3 µg/kg), etomidate (0.2–0.3 mg/kg), and atracurium (0.6 mg/kg) were used to induce general anesthesia. Following intubation, sevoflurane (1.5-3%) was inhaled to maintain BIS values between 40 and 60. Intermittent atracurium injections kept muscles relaxed, and the continuous infusion volume of remifentanil (0.1–0.2 µg/(kg·min)) was adjusted to maintain the heart rate (HR) and the mean arterial blood pressure (MAP) within 80–120% of baseline. When the MAP was less than 80% of the baseline value, IV ephedrine or IV phenylephrine was administered, and IV atropine was administered when the HR was less than 40 bpm. IV 5 mg dexamethasone were administered for PONV prevention. Flurbiprofen (100 mg) was injected intravenously 30 min before the end of the operation to ensure postoperative analgesia was achieved. When the patient was awake and the tidal volume was sufficient, the tracheal tube was removed. The anesthesiologist then transferred the patient to the postanesthesia care unit (PACU).

After induction, the patient was changed to a prone position. The skin was disinfected with iodophor, and the high-frequency linear ultrasound probe (6–13 Mhz, SonoSite S Series, USA) was placed in a sterile sleeve. The probe was placed longitudinally at the 12th rib of the midscapular line and moved inward along the 12th rib until the tip of the transverse process was reached. Using an in-plane method, the needle (0.71 × 120 mm, 22 G, B Braun, Germany) was inserted into the skin at the head of the probe until it reached the transverse process. Two milliliters of normal saline (NaCl 0.9%) were injected to identify the needle tip position. If the needle tip was located between the transverse process and the ESM and was not in the muscle, 20 mL 0.4% ropivacaine (AstraZeneca AB, Sweden) was injected immediately (Fig. 1). The procedure was then repeated on the opposite side.

Ultrasound images of ESPB with the in-plane technique. Arrows, needle shaft. Abbreviations: ESM, erector spinae muscle; TP, transverse process

Both groups were administered intravenously guttae flurbiprofen 100 mg twice a day to maintain postoperative analgesia. When the numeric rating scale (NRS) score was ≥ 4, an intramuscular injection of 100 mg tramadol was administered.

​ Intraoperative data were collected, including the duration of surgery, remifentanil use, total blood loss, hemodynamic parameters, and vasoactive medication dosage. Postoperative data included extubation time, Sedation-Agitation Scale (SAS) score [8] after extubation, NRS score (evaluated at 0, 0.5, 4, 12, 24, 36, 48, 60, and 72 h postoperatively), tramadol use within 72 h postoperatively, incidence of POD within 3 days after surgery (POD diagnosed by Confusion Assessment Method (CAM)) [9], postoperative nausea and vomiting (PONV), ESPB-related complications, ambulation time (the period between the completion of surgery until and when the patient could stand up and walk with support from a family member), and length of hospitalization after surgery.

Intraoperative data were collected by the anesthesiologists and postoperative data were obtained by the trained investigators.

The primary outcome was the NRS pain score (0–10; 0, no pain; 1–3, mild pain; 4–6, moderate pain; 7–10, severe pain) at 12 h after surgery. Secondary outcomes included the NRS pain score and tramadol use within 72 h postoperatively, intraoperative remifentanil use, hemodynamic parameters, extubation time, SAS score after extubation, incidence of POD, incidence of PONV, ESPB-related complications, ambulation time, and length of hospitalization after surgery.

The sample size was estimated with the MedSci Application (V.6.2.2, MedSci, China). We performed sample size calculations, in which a 1-point reduction in the NRS score at 12 h was considered clinically relevant. Eight patients in each group were included in our pilot study. The NRS score was 2 (1) (mean (SD)) in the ESPB group and 2.5 (1.4) in the control group at 12 h. According to a two-sided unpaired t test, 90 patients were required in each group to achieve 80% response and an α threshold of 0.05. Due to the nonparametric estimation parameter data, the sample size was increased by 15%, resulting in 104 patients per group. Considering loss to follow-up, we selected 115 patients per group. Statistical tests were conducted with IBM SPSS Statistics 26.0 (IBM, USA). The Shapiro―Wilk test was employed to test for normal distribution; the Student’s t test and the Mann―Whitney U test were used for continuous variables (normal and nonnormal distribution);and the Fisher’s exact or χ2 tests were used for categorical variables and proportions. A two-tailed, P value < 0.05 was deemed statistically significant.