This prospective, single-centre, randomized, parallel-group, single-blinded trial (Chinese Clinical Trial Registry: ChiCTR 1800015141) was approved by the Ethics Committee of Harbin Medical University. The study adhered to CONSORT guidelines and informed written consent was obtained from all patients.
After institutional review board approval (Harbin Medical University Institutional Research Board: KY2018-003), 150 ASA II-III Chinese patients aged 20-70 who underwent open splenectomy were included in this trial between March 2018 and July 2018. Patients with an ASA physical status of 4 or more, an allergy to local anaesthetics, a history of abdominal surgery, a body mass index < 15 kg.m-2 or > 40 kg.m-2 or severe cardiac and/or pulmonary dysfunction were excluded. Patients with acute or chronic preoperative opioid consumption any other analgesic treatment for chronic pain before surgery, or psychiatric or neurological factors (language barrier, neuropsychiatric disorder) were excluded. Patients who required postoperative mechanical ventilation, had sustained excessive haemorrhage (>1 L of estimated blood loss) or required a massive transfusion and patients with failed nerve block (the needle could not be positioned to the anatomic structure and the drugs failed to enter the interspace) were also excluded.
These 150 patients for whom TAPB and RSB were successfully established were randomly divided into 3 groups: a control group (C), a levobupivacaine group (L) and a levobupivacaine/morphine group (LM) (n=50). Patients in the group C received general anaesthesia combined with RSB and TAPB with saline, and intravenous PCA for postoperative pain, and patients in the group L and group LM received general anaesthesia combined with RSB and TAPB with 0.2% levobupivacaine alone or levobupivacaine with morphine 30 mg.kg-1. The dosage of morphine was adjusted according to the paravertebral block .
All patients were monitored by continuous electrocardiography (ECG) and pulse oximetry (SpO2). After local infiltration of lidocaine, the radial artery was intubated to monitor the invasive blood pressure (BP). After induction with 0.03 mg.kg-1 midazolam, 1 mg.kg-1 lidocaine, 0.4 mg.kg-1 sufentanil, 0.5 mg.kg-1 atracurium, and 0.2 mg.kg-1 etomidate, the trachea was intubated. After intubation, the patients were randomized into the groups C, L and LM. Patients in the group C received intravenous PCA with sufentanil (0.04 mg.kg-1h-1) diluted into 150 ml of saline with a PCA device at a rate of 2 ml.h-1 continuously, a 2-ml bolus injection, and PCA with a 15-min lockout interval for postoperative analgesia. Patients in groups L and LM received levobupivacaine (0.2% levobupivacaine diluted into 60 ml of saline) or levobupivacaine combined with morphine (0.2% levobupivacaine and morphine 30 mg.kg-1 diluted into 60 ml saline) for postoperative analgesia. Anaesthesia was maintained with sevoflurane (expiratory concentration 1.5%) and remifentanil. Patients in group C received remifentanil (10 mg.kg-1h-1), and patients in group L and group LM received remifentanil to maintain the BP and heart rate within from 20% of the baseline. If the change in BP and/or heart rate (HR) exceeded 20% of baseline, 1 mg.kg-1 remifentanil or 6 mg ephedrine was injected.
The randomization of patients allocated to the group C, L or LM was in accordance with a random sequence generated using Stata_version 11 software (StataCorp; TX, USA). An independent anaesthesiologist who did not participate in the peri-operative evaluation prepared the drug for each group according to a code known only to the anaesthesiologist. Another anaesthesiologist who was blinded to the randomization and anaesthesia investigated and recorded the peri-operative data.
All patients received RSB and TAPB after intubation in a supine position.
Ultrasound-guided rectus sheath block
An incision was made from the subxiphoid to the anterior axillary along the left subcostal margin. We performed RSB at the first and second segments of the rectus abdominis muscle. Briefly, under ultrasound guidance (M-Turbo- Ultrasound system; SonoSite, Bothell, WA, USA), a 38-mm broadband linear array ultrasound probe (5–10 MHz) was positioned at the level of the first segment of the subxiphoid and the next segment lateral to the left of the longitudinal plane to the abdominal wall. After confirmation of the rectus abdominis muscle, the 24 G needle was inserted into the skin under the middle of the ultrasound probe and then pierced into the posterior rectus sheath. Saline was injected to confirm the placement of the needle at the posterior rectus sheath. When the needle placement into the rectus sheath was confirmed, 15 ml of saline, 0.2% levobupivacaine or 0.2% levobupivacaine combined with morphine was injected for the 3 groups after confirmation that no blood was withdrawn, leading to the appearance of a hypoechoic space. Then, RSB of the next segment was performed using the same method and the same volume of anaesthetic solution (Fig. 1a and b).
Ultrasound-guided transversus abdominis plane block
The ultrasound probe (5-10 MHz) was positioned in the left rectus abdominis muscle laterally between the subcostal margin and the iliac crest to obtain the classical image of abdominal layers, including the external oblique muscle, the internal oblique muscle, the transversus abdominis muscle, and the peritoneum. To obtain a large blockade area, we continuously injected anaesthetics during the insertion of the needle from the beginning of the transverse fascia (rectus abdominis muscle laterally) to the midaxillary line. When the tip of the needle had been advanced to the beginning of the transabdominal plane, the needle was inserted along the transabdominal plane, and regional anaesthetics were injected step-by-step to assure that the whole transabdominal plane was filled with anaesthetics (Fig. 1c, d and e).
procedures and measurements
Blood was collected at completion of the TAPB and then at 10, 20, 30, 60, 90, 120 and 150 min after injection of local anaesthetics to measure the concentration of levobupivacaine using high-performance liquid chromatography (HPLC). Briefly, plasma was collected by centrifugation at 3000 rpm.min-1 for 10 min and kept frozen at -20℃ for subsequent HPLC (CBM-20A HPLC, Kyoto, Japan) test. The sample flow rate was set to 1.0 ml.min-1and the detection wavelength was 210 nm. The levobupivacaine concentration was calculated according to the concentration curve of levobupivacaine hydrochloride (Hengrui, Jiangsu, China). The calculated curve of levobupivacaine showed good linearity over a range of 0.5-2000 ng.ml-1 (correlation coefficient ≥0.99).
After RSB and TAPB, the right subclavian vein was cannulated to collect blood and infuse blood or fluids, and all patients underwent standard open splenectomy . To avoid the influence of the surgical procedure on postoperative pain, all enrolled patients received RSB and TAPB by the same surgery team. All patients received a left subcostal incision in the supine position. Postoperative analgesia was provided with PCA using sufentanil (0.04 mg.kg-1h-1) in the C group. The patients in the nerve block groups received an injection of morphine when the pain score exceeded 4. All patients received 40 μg.kg-1 granisetron to prevent PONV . After extubation, all patients were transferred to the postoperative care unit (PACU). When the SpO2 was over 95% in air, the patient was transferred to the ward.
Blood loss, infusion (red blood cells [RBCs] and plasma) and consumption of remifentanil were recorded. Postoperative pain at rest and upon coughing was evaluated with a visual analogue scale (VAS) at 0, 2, 4, 6, 24, 48 and 72 hours after operation. The postoperative pain was evaluated by incision and visceral pain (0=no pain, 10=worst pain). Other variables were recorded, including time to first exhaust, time to first defecation, time to first oral intake, time to first off-bed activity and incidence of PONV (scaled from 0 to 10). Before discharge, all patients scored their satisfaction with postoperative analgesia (poor=0; fair=1; good=2; excellent=3).
If the VAS score was greater than 4, a 3 mg intravenous (i.v.) bolus of morphine was administered, and pain was reassessed after 10-15 min . Metoclopramide (10 mg) was intravenously injected if the patients reported a single episode of severe nausea (>7) or any episode of vomiting. The primary outcome was the consumption of analgesics over 24 hours. The pain score, sedation score, satisfaction score postoperative recovery time and PONV were secondary outcomes. To guarantee objective results, the investigator was blinded to the randomization and anaesthesia.
In our preliminary pilot study and our own experience, the consumption of morphine during the postoperative 72 hours was approximately 15.8 (6.4) mg in patients without any other analgesics. Approximately 46 patients in each group were required to detect a 25% reduction in morphine between the control and LM groups and to achieve 80% power with a two-sided alpha of 0.05.
The normality of the data was analysed with the Shapiro-Wilk test. Normally distributed data are presented as the mean (SD). Non-normally distributed data are presented as the median [interquartile range (IQR)]. Continuous data were analysed with repeated measures analysis. Normally distributed data were analysed with a t test, and non-normally distributed data were analysed with the Mann-Whitney test. Categorical data were analysed with the chi-square test.