Single-shot Pectoral Plane(PECs) Block Versus Continuous Local Anaesthetic Infusion Analgesia or Both PECS Block and Local Anaesthetic Infusion After Breast Surgery: A Prospective Randomised, Double-blind Trial

Methods The Mater Misericordiae University Hospital's Institutional Review Board (IRB) approved this study; reference number 1/378/1840, dated December 21st, 2016. It was registered with clinicaltrials.gov; reference number NCT03024697, dated February 15th, 2017. Forty-five ASA I-III female patients aged 18 years or older, undergoing non-ambulatory breast cancer surgery under general anaesthesia between January and May 2017 were included. All participants provided informed written consent.

Non-ambulatory breast cancer surgery encompassed wide-local excision with lymph node resection, simple mastectomy, and mastectomy with lymph node resection. Those undergoing latissimus dorsi or DIEP (deep inferior epigastric perforator) flap reconstruction were excluded. Patients with chronic pain syndromes, local anaesthetic allergy, contraindication to simple analgesics, local infection over the proposed block site, coagulopathy, or co-morbid conditions precluding the provision of informed consent, were excluded. All patients were day of surgery admission and attended a preoperative anaesthetic assessment clinic.

Patients were allocated into three groups using computer-generated randomisation, with the study number and group allocation concealed in sealed envelopes. Blocked randomisation in groups of 9 was applied, giving similar numbers in each group as the study progressed. Groups were named 'PECs', 'Local anaesthetic infusion (LA infusion)' , and 'Both (PECS \& LA infusion)', which were evenly distributed to opaque envelopes numbered one to forty-five in accordance with randomisation. The randomisation key was held by an independent party, and was not used to reveal participant group allocations until data analysis commenced. 'PECs' patients received PECS I \& II blocks and a sham wound infiltration catheter. 'LA infusion' patients received a continuous local anaesthetic wound infusion catheter. 'PECs \& LA infusion' patients received a combination of techniques, without the initial local anaesthetic bolus via the wound infusion catheter. PECs blocks were performed while patients were under general anaesthesia, prior to the commencement of surgery. The operative surgeon sited wound infusion catheters during skin closure at the end of surgery.

Patients were induced with fentanyl 1-2 μg.kg-1, followed by propofol titrated to the absence of verbal response. Anaesthesia was maintained using an oxygen, air and sevoflurane combination. Airway management and lung ventilation strategies were at the discretion of the supervising anaesthetist. Patients received a standard intraoperative analgesic regimen of paracetamol 1g and dexketoprofen 50mg intravenously (IV), with rescue morphine as deemed necessary. Combination antiemetics, including ondansetron 0.1-0.15 mg.kg-1 IV, dexamethasone 0.1-0.2 mg.kg-1 IV or droperidol 0.01-0.015 mg.kg-1 IV, were administered according to patient risk factors. Intraoperative management was otherwise left to the discretion of the supervising anaesthetist. Routine monitoring was used in accordance with AAGBI guidelines15. An electronic anaesthetic record was used to document physiological parameters. Perioperative events, such as induction of anaesthesia, initial skin incision, and end of surgery, were annotated on the record.

PECs I \& II blocks were performed on the side of surgery, using the ultrasound-guided technique described by Blanco et al9, 10. The patient was placed in the supine, head-up, position with the arm abducted. The skin was prepared with chlorhexidine gluconate 2%/isopropyl alcohol 70% (BD, ChloraPrep, NJ, USA). PECs blocks were performed with a 22-gauge echogenic needle (B. Braun, Ultraplex 360 cannula, Hessen, Germany; 50-80mm), using the same ultrasound machine (Sonosite Edge, Sonosite, Inc., Bothwell, WA, USA) and transducer (Sonosite HFL 50x, Sonosite, Inc., Bothwell, WA, USA). The ultrasound probe was placed inferolaterally at the mid-clavicular level. The axillary artery and vein were identified, and the probe moved laterally until pectoralis major, pectoralis minor and serratus anterior muscles were located at the level of the 3rd rib. A needle in-plane approach was taken until the needle tip was positioned in the plane between pectoralis major and minor muscles, and levobupivacaine 0.25% 10ml was injected. The needle was advanced until it occupied the space between pectoralis minor and serratus anterior muscles, and a further 20ml of levobupivacaine 0.25% was injected.

Blinding was not considered necessary for 'LA infusion' group patients, as PECs blocks were performed under general anaesthesia. All patients had a wound infusion catheter sited and attached to an elastometric pump (B. Braun, ON-Q PainBuster, Hessen, Germany), which was placed by the operating surgeon during skin closure at the end of surgery. For blinding purposes, patients in the 'PECs' group received a sham wound infiltration catheter system. 'LA infusion' and 'PECS \& LA infusion' patients received levobupivacaine 0.1% at 10ml.hr-1 for 24 hours postoperatively, while 'PECs' patients received sodium chloride 0.9% at the same rate. 'LA infusion' patients were administered an initial bolus of levobupivacaine 0.25% 20ml at the end of surgery. Those who performed the procedures, or were involved in the perioperative management, were not involved in postoperative pain assessment or data collection. Study interventions as described occurred while patients were under general anaesthesia.

Patients were monitored for 24 hours after surgery, initially in the post-anaesthesia care unit (PACU), and then at ward level once PACU discharge criteria were met. Oxycodone 1-2mg IV as required was prescribed for rescue analgesia in the PACU. A standard analgesic protocol of paracetamol 1g PO/IV 6-hourly regular, ibuprofen 400mg PO 8-hourly regular, and oxycodone immediate release (oxynorm) 5-10mg 1-hourly as required for rescue analgesia, was prescribed for all patients. Ondansetron 4-8mg IV 8-hourly as required was prescribed, to be administered in the event of postoperative nausea and vomiting (PONV).

The primary outcome measure of the study was area under curve of verbal rating score (VRS) of pain moving versus time, (AUC) where VRS pain was measured at 1 hour, 4-6 hours, 10-14 hours, and 20-24 hours postoperatively. At each assessment, patients were invited to report their pain VRS pain at rest. Pain while moving was defined as pain experienced while sitting forward from a recumbent position. Secondary outcome measures were total opioid consumption over a 24-hour period and presence of adverse events. An investigator masked to group allocations recorded vital signs, pain scores, antiemetic administration and opioid use postoperatively. Adverse events such as sedation, respiratory depression, hypotension, pruritus and PONV were also recorded.

Statistical Analysis

Data was analysed by Graph Pad Prism v6 (Salt Lake City, UH). The primary end-point was area under the VRS moving versus time curve (AUC). We powered the study to detect a 25% difference in AUC from an internal pilot study which indicated mean AUC of 60 cm.hr-1 (SD 5-10) among patients receiving PECs or LA infusion in our clinical service, to 45 cm.hr-1. We had observed the standard deviation of the AUC to be in the order of 15 cm.hr-1. Assuming a Type I error of 0.05 and a Type II error of 0.1, then n=11 patients would be required each group to detect this difference with 90% power. We obtained IRB permission for n=15 each group to allow for protocol violations or patient drop-out.

Data was collected at 1, 4-6, 12-14 and 20-24 hr, post-operatively. It was tested for distribution initially using the Kolmogorov-Smirnov test. For each patient, their AUC both moving and at rest was calculated by plotting their pain VRS versus time using Graph Pad Prism v.6. The mean AUC according to the three study groups was then calculated and inspected for distribution as described above, and all AUC data was found to be normally distributed. Therefore, AUC data is expressed as mean ± SD and was compared using ANOVA with post hoc Bonferroni correction. Non-normally distributed data was compared using the Kruskal-Wallis test for repeated measurements. Categorical data (e.g. the incidence of potential adverse effects) were compared using Fischer's exact test where appropriate. P<0.05 was deemed statistically significant.