Our multi-centre randomized, controlled, double-blinded study was approved by the Scientific and Ethics Committee of the Centre intégré universitaire de santé.

et de services sociaux de l’Est-de-l’Île de Montréal associated with the Hôpital Maisonneuve-Rosemont and registered at clinicaltrials.gov (NCT03860480 registred on 21/01/2019). After providing written informed consent, patients undergoing elective VATS at the Hôpital Maisonneuve-Rosemont and the Centre hospitalier de l’Université de Montréal (CHUM) tertiary university hospitals were enrolled from November 2018 to December 2021. We randomized these patients into two groups: bupivacaine 0.5% 30 ml with epinephrine 5 mcg ml− 1 (Bupivacaine group) or 30 ml of normal saline (Control group).

Indication for surgery was an elective pulmonary wedge resection, segmentectomy, or lobectomy. All patients were above 18 years old and had an American Society of Anesthesiologists physical status score of I to III. Exclusion criteria were: BMI > 35 kg m− 2, chronic pain with regular use of either opioids or gabapentinoids during the 2 weeks before surgery, regular marijuana use, history of thoracic surgery on the same side, anticipated high risk of conversion to thoracotomy, inability to communicate with the investigators, taking anticoagulation or antiplatelet medications (except Aspirin and nonsteroidal anti-inflammatory drugs), suffering from any bleeding disorders, surgery for empyema and sympathectomy, known allergy to local anesthetics or hydromorphone or fentanyl, active infection at the injection site, pre-existing neurological or psychiatric illness, severe cardiovascular disease, liver failure, renal failure (estimated glomerular filtration rate less than 15 ml min− 1 per 1.73 m2 ), and pregnancy. Patients were also excluded after randomization if they had a perioperative conversion to thoracotomy, severe intra- or postoperative bleeding, required postoperative mechanical ventilation, or a technical failure to proceed with the blocks.

All patients were contacted with a pre-approved script by phone, at our preoperative investigation clinic, or on the day of surgery by one of our attending physicians, an anesthesiology resident, or a research nurse. On the day of the surgery, we obtained written informed consent with written information provided and randomized patients into two groups using computer-generated random numbers and a 1:1 allocation ratio. We completed preoperative QoR-15 questionnaires for all patients. They all received pre-emptive 975 mg oral acetaminophen. A research nurse, otherwise not involved in the care of the randomized patient, prepared the research medication. Identical syringes were used for both groups, blinding the surgeon, the attending anesthesiologist, the research investigators, the nursing staff, and the patient. We maintained blinding throughout the study.

We monitored all patients using a 5-lead electrocardiogram, non-invasive blood pressure, pulse oximetry, end-tidal carbon dioxide, train-of-four stimulation, and a bispectral index device (BIS).

General anesthesia protocol was the same for all patients. We administered all anesthesia and analgesia drugs following an estimation of the adjusted body weight using the Schwartz S.N. formula and followed a standard preoperative fasting protocol. After a 3-min preoxygenation period, we achieved a balanced induction using fentanyl 2 mcg kg− 1, propofol 1 to 3 mg kg− 1, remifentanil 0 to 1 mcg kg− 1, and rocuronium 0.6 to 1 mg kg− 1. We used remifentanil to blunt the hemodynamic response to intubation and fentanyl for analgesia during the procedure in relay to remifentanil. At the beginning of the procedure, each patient received standard pre-emptive postoperative nausea and vomiting (PONV) prevention comprising dexamethasone 4 mg IV and ondansetron 4 mg IV at the end.

We intubated and mechanically ventilated all patients using volume-controlled positive-pressure ventilation with a tidal volume of 4–6 ml kg− 1 of adjusted weight to maintain end-tidal carbon dioxide tension at 35–45 mm Hg. The lung isolation technique was standardized for all patients with a left double-lumen endotracheal tube.

We maintained anesthesia using sevoflurane and aimed for a BIS of 40 to 60 to a maximum of 1.0 MAC adjusted for the patient’s age with a minimal FiO2 to sustain a saturation of 90% or greater. We administered fentanyl doses of 25 mcg IV to maintain blood pressure and heart rate within 20% of baseline values. We paralyzed patients using rocuronium for the surgery and used a train-of-four stimulation throughout the case.

We extubated the patients before transferring them to the post-anesthesia care unit (PACU). The PACU nursing team administered hydromorphone 0.4 mg IV boluses as needed via the PCA pump every 5 min up to 1.2 mg if the pain was more than mild (VNRS score of 4 or more). After that, we instructed patients to use a PCA pump as needed. All hydromorphone doses received were recorded by the PCA pump. PCA settings for both groups were hydromorphone with a bolus of 0.2 mg, lockout time of 5 min, and no background infusion.

After placing the patient under GA in a standard lateral position, an ESP block expert (≥ 20 ESP blocks completed) performed an ESP block at the T5 transverse process using an in-plane approach. ESP block experts performed all blocks in our study. They placed a high-frequency linear probe covered in a sterile sheath in the longitudinal plane 3 cm parasagittal from the midline on the operated side, as defined by Forero et al. 2016 [8]. After standard chlorhexidine disinfection and identification of the transverse process and the trapezius, rhomboid, and erector spinae muscles, the expert inserted an 18 Ga Contiplex Ultra 360 needle in a cephalad-caudad direction until the tip reached the fascial plane deep to the erector spinae muscle. Hydrodissection with normal saline was used to correct and confirm the needle tip position. After confirmation, the expert administered an unidentified solution based on the randomized group and inserted a 20 Ga polyamide catheter up to 3 cm in the identified plane. In the event of conversion to thoracotomy, we excluded the patient from the study and used the catheter postoperatively for up to 4 days with continuous bupivacaine perfusion. If the procedure went as planned, we removed the catheter at the end of the surgical procedure.

Our primary outcome was the total hydromorphone consumption at 24 h. Secondary outcomes included hydromorphone consumption at 1, 6, 12, and 18 h post-PACU arrival, thorax and shoulder VNRS score at rest and during cough at 1, 6, 12, 18, and 24 h post-PACU arrival, QoR-15 on postoperative-day 1, Ramsay Sedation Scale, pruritus and PONV at the same time points.

Registered nurses recorded total hydromorphone consumption readings from the PCA pumps at different time points after surgery. They also recorded VNRS scores for thorax and shoulder pain at rest and during cough at 1, 6, 12, 18, and 24 h post-PACU arrival and evaluated the PONV score, the Ramsay Sedation Scale, the incidence of PONV, and the incidence of pruritus at the same time points.

The QoR-15 is a patient-reported questionnaire with 15 questions assessing the quality of recovery from anesthesia and surgery. Each question is rated on a scale of 0 to 10, with the highest score being 150 and the lowest being 0. As the score approaches 150, the patient’s recovery is considered improved. The questionnaire evaluates five elements: pain intensity, physical well-being, autonomy, and psychological and emotional states [9]. An anesthesiologist, anesthesiology resident, or research nurse administered the French version of this questionnaire.

Based on a preliminary study at the Hôpital Maisonneuve-Rosemont, the mean consumption of IV hydromorphone was 4.7 (3) mg in the first 24 h after VATS lung resection in the Control group. Therefore, using a two-tailed alpha of 0.05, we calculated that a sample size of 52 patients (26 per group) would provide 80% power to detect a clinically significant decrease of 50% in hydromorphone consumption at 24 h in the Bupivacaine group.

We performed statistical analysis using Prism 7 for Mac OS X, version 7.0d (GraphPad Software, Inc.). The normality of the distribution of hydromorphone consumption was confirmed using the D’Agostino & Pearson normality test. We compared the 24-h total doses of hydromorphone and the consumption of hydromorphone at 1, 6, 12, 18, and 24 h post-surgery with unpaired t-test and expressed as mean (SD). Ramsay Sedation Scale and VNRS and PONV scores were compared using the two-tailed Mann-Whitney test and expressed as median (IQR). Using Fisher’s exact test, we compared the incidence of PONV and pruritus between groups. We used post-hoc repeated measures analysis of variance (ANOVA) to evaluate the effect of the randomized group, time, and health care centre where the surgery was performed (independent variables) on hydromorphone consumption and QoR scores (dependent variables). We also evaluated independent variables for interaction. To control for type-I errors at 5%, we used the Bonferroni correction and a corrected P-value for the secondary outcomes. A P-value ≤ 0.05 was considered statistically significant.