This was a randomized controlled pilot trial that compared two parallel groups for superiority of intervention versus placebo. Ethical approval for this study (Ethics Committee No. P2016/404/2016-003918-27) was provided by the Ethics Committee Erasme Hospital, 808 route de Lennik, B-1070 Brussels, Belgium (Chairperson Prof J.-M. Boeynaems) on 24 October 2016. All included patients signed a written informed consent before participation. This study adheres to CONSORT guidelines for reporting clinical trials (see additional file: « CONSORT Checklist »).
Adult (>18 years old) patients scheduled for elective major abdominal surgery were investigated for eligibility. Patients scheduled for colonic or bariatric surgery were considered to have a ‘moderate risk’, and thus not included in the eligibility screening. Patients scheduled for hepatic resection were excluded to prevent potential accumulation of lidocaine because of unpredictable changes in its pharmacokinetics. Patients who were to be managed with combined epidural and general anesthesia (e.g. esophagectomy) were excluded because of potential parallel administration of another LA. Patients presenting one or more of the following medical conditions were excluded because of increased risk of lidocaine accumulation and/or local anesthetics systemic toxicity: severe heart conduction blocks without implantable pacemaker, severe liver and kidney insufficiency (Kidney Disease: Improving Global Outcomes [KDIGO] stage >3a), acute heart failure, and known allergic reactions to any amide-linked LAs. Patients with atrial fibrillation were also excluded because it was impossible to follow the fluid administration protocol (discussed in the ‘Interventions’ section).
This study was conducted at Erasme Hospital (Brussels, Belgium), a tertiary healthcare institution of the Université Libre de Bruxelles. Patients were enrolled from December 2016 to March 2017.
Patients allocated to the intervention (LIDO) group received 1.5 mg kg-1 (total body weight [TBW]) of 1% lidocaine (Xylocaïne®; AstraZeneca, Cambridge, United Kingdom) just before anesthesia induction, which was immediately followed by a 2 mg kg-1 h-1 (TBW) continuous intravenous infusion until skin closure. Patients allocated to the placebo (PLA) group received an equivalent volume of 0.9% saline (0.15 ml kg-1 bolus and 0.2 ml kg-1 h-1 continuous intravenous infusion). Anesthesia protocol was standardized for both groups. When indicated, an intrathecal injection of 0.1–0.3 mg of morphine was administered before induction. The latter was achieved using propofol, remifentanil (with target-controlled infusion [TCI]) and a neuromuscular blocking agent (usually rocuronium or cisatracurium). Every patient received dexamethasone (10 mg). Maintenance was achieved using desflurane and remifentanil (TCI), guided to maintain the Bispectral Index (BIS™; Aspect Medical Systems, Norwood, MA, USA) readings between 40 and 60 with 0% of burst suppression rate. Hemodynamics were managed with a goal-directed therapy (GDT) protocol, based on stroke volume variation (SVV, measured using the FloTrac™ system [EV1000; Edwards Lifesciences Corp., Irvine, CA, USA]) and mean arterial pressure (MAP). Plasmalyte® (Baxter, Deerfield, IL) administered at 2 ml kg-1 h-1 was the basal infusion. Triggers for fluid bolus administration (250 ml of crystalloids in 10 min) and for catecholamine optimization were SVV≥13% and MAP<70 mmHg (Fig. 1), respectively. When to transfuse hemoderivates and administer colloid solutions and the management of postoperative analgesia were at the discretion of the anesthetist in charge of the patient.
Endothelial function and the EG were investigated using three techniques: concentration of syndecan-1, measurement of the tissue oxygen saturation (StO2) during the vascular occlusion test (VOT), and contemporary measurement of flow-mediated dilation (FMD). Data and blood samples were collected before surgery (T0), at 1–3 h post-surgery in the recovery room (T1) and 24 h post-surgery in the surgical ward (T2).
The concentration of syndecan-1, a marker of EG shedding, was measured by classic sandwich enzyme-linked immunosorbent assay (ELISA), based on the manufacturer’s instructions (Syndecan-1 ELISA kit; Tebu-bio, Boechout, Belgium). Blood samples were collected in dry tubes, centrifuged to obtain the serum and stored at -80°C for a maximum of 6 months before final analysis.
The StO2 was measured continuously (every 2 seconds) and noninvasively using near-infrared spectroscopy (NIRS) (ForeSight®; CASMED®; CAS Medical Systems, Inc., Branford, CT, USA) on the thenar eminence, as previously described.11 At the same time, the diameter of the brachial artery and flow velocity were measured continuously by using a 5-12MHz linear ultrasonography transducer (Sparq®; Phillips, Amsterdam, the Netherlands), which was applied to the upper arm with mechanical support for image stabilisation (Image 1 in « Additional File 1 » shows the standard set up for a participant). Baseline values were acquired over 1 minute. A VOT was then conducted by rapidly inflating a pneumatic cuff, which was placed around the forearm, up to 200 mmHg (or 50 mmHg suprasystolic pressure). After 5 minutes the cuff was released, and the hyperemic response was evaluated for another 4 minutes. The following StO2-derived variables were analysed: StO2-baseline, StO2-ischemic slope and StO2-reperfusion slope (Image 2 in « Additional File 2 » shows the evolution of StO2 during the test in one participant). The FMD was assessed by automated edge detection software (FMD Studio™ (CardioVascular Suite™); Quipu srl, Pisa, Italy),19 based on the experts’ guidelines.20
The following FMD-derived variables were analysed: brachial artery baseline diameter, FMD-maximum (i.e. the maximal diameter during reperfusion) and the area under the curve of estimated shear rate of hyperaemic flow until FMD-max (Image 3 in « Additional File 3 » shows the evolution of the brachial artery diameter and shear rate during a test in one participant).
Data concerning fluid requirements were prospectively collected during surgery and during recovery in the postanesthesia care unit (PACU). Haemodynamic variables were collected only during surgery by the EV1000 clinical platform.
The primary endpoint was the evolution of the syndecan-1 concentration postoperatively in the LIDO group, compared with the PLA group (hereafter referred to as ‘difference between groups’). Predefined secondary outcomes were the effect of lidocaine on NIRS- and FMD-derived variables (i.e. difference between groups); the influence of surgery on NIRS and FMD-derived variables and its association with group assignment (hereafter referred to as ‘difference between times’); the correlation between glycocalyx, microcirculation and vascular reactivity at three time points; and the influence of group assignment on fluid requirements. Potential harmful effects of lidocaine were also systematically researched and reported. Haemodynamic variables, even if not originally included in secondary endpoints, were also taken in account for analysis and validation of compliance to the GDT protocol.
We were unable to find any published study on the effects of lidocaine on endothelial function in the clinical setting, and evidence concerning alteration of EG in major abdominal surgery is scarce. Hence, we decided to perform a pilot study which included 40 patients with 20 patients for each group.
Participants were randomly assigned to one of two groups in a 1:1 ratio, based on Efron’s biased coin randomisation procedure generated with NCSS v10 Statistical Software (2015, NCSS, Llc. Kaysville, UT, USA).
Patients, healthcare providers, data collectors and outcome adjudicators were all blinded to group assignment. The physician in charge for generation of allocation sequence and concealment was not directly implicated in treatment administration or data collection.
Data are presented as the mean ± the standard deviation. Data were compared between the groups using the Mann–Whitney test or by two-way analysis of variance (ANOVA) for repeated measures, as indicated. One factor was the study group and a second factor was time. For each variable, the three null hypotheses of the two-way ANOVA tests were that the means of the observations grouped by one factor would be the same; that the means of the observations grouped by the other factor would be the same; and that there would be no interaction between the two factors. The P value would be indicated for the difference between groups (i.e., all time points together) or for the difference between times (i.e., all groups together), and for the interaction between groups and times. A Tukey-Kramer multi comparison test was performed to examine all pairs of treatment means. The other continuous non longitudinal variables were compared with the Mann-Whitney test. For all tests, P<0.05 was statistically significant. These computations were performed using the software package Systat version 5.0 for DOS (Systat, Inc., Evanston, IL, USA).