Immunological effects of post-operative epidural analgesia versus oral opioids in VATS

Surgery remains the primary treatment for many malignant tumours. However, some primary cancer cells may persist after surgery, and tumours might dislodge and spread via the lymphovascular system. Whether these cells proliferate or lead to metastases is generally believed to depend on the balance between tumour aggressiveness and the body's capability to suppress the tumour load [1]. Surgery and trauma induce immunosuppression by an initial pro-inflammatory immune response mediated primarily by the cells of the innate immune system, followed by a compensatory anti-inflammatory reaction mainly by the adaptive immune system, restoring homeostasis [2]. Studies suggest that the anaesthetics and the adjuvants used in the perioperative period may possibly affect cancer recurrence and survival [3]. Though some argue that it remains speculative whether any aspects of anaesthetic management influence cancer outcomes [4], a reasonable prerequisite would be that a detectable impact on the immune system is present. In the present study, we investigated the effect of epidural analgesia on the innate, non-specific immune system in patients undergoing videoscopic surgery for lung cancer by examining selected cytokines and post-operative survival, testing the hypothesis that epidural analgesia may attenuate post-operative immunosuppression.

METHODS

Study design and participants

We performed a patient- and observer-blinded, parallel-arm, randomised controlled trial at a single cardiothoracic anaesthesia and surgery department in a tertiary care hospital in Denmark, examining various aspects of using supplementary epidural analgesia in video-assisted thoracoscopic surgery (VATS) for lung cancer (ClinicalTrials NCT02359175). This trial was an adjunct trial focusing on the immunological aspects of epidural anaesthesia. Patients scheduled for elective VATS lobectomy were eligible for trial participation and screened for inclusion. If the inclusion criteria were met, we obtained written informed consent and enrolled the patient. The exclusion criteria were age < 18 years, pregnancy, contraindications to epidural catheter insertion, pre-existing other cancers or immunodeficiency, previous thoracic surgery or daily use of analgesics. We used a two-arm treatment design, with one group receiving thoracic epidural analgesia for post-operative pain management (TE Group). In contrast, the other received oral morphine (OM Group). Both groups received blinded baseline analgesics. Baseline medication, administered as premedication before surgery and at fixed predefined intervals after surgery, comprised oral acetaminophen (1 g), ibuprofen (400 mg) and placebo in the TE Group, while the OM Group received acetaminophen, ibuprofen and morphine (10 mg). Both groups received identical intraoperative pain management regimes that were replaced post-operatively with the allocated epidural infusion of either bupivacaine 1.0 mg/ml with fentanyl 2.0 μg/ml (TE Group) or saline (OM Group).

The trial study was approved by The Regional Scientific Ethical Committees for Southern Denmark, The Danish Health and Medicines Authority, and The Danish Data Protection Agency. The trial protocol and dataset are available upon request.

Randomisation and masking

According to a computer-generated randomisation list, the hospital pharmacy provided similar, consecutively numbered pre-packed boxes with the study medication. The list was generated from a validated source using block randomisation with blocks of four and a treatment-to-placebo allocation of 1:1 without stratification.

Study procedures

All participating patients were seen before surgery to assess operability and get an immunological baseline blood sample. The study procedure has been described in detail previously [5]. In short, after epidural catheter insertion and test of function, we allocated patients using a consecutive randomisation number and administered premedication. General anaesthesia was induced and maintained using propofol and remifentanil in conjunction with an epidural infusion of bupivacaine (2.5 mg/ml) containing sufentanil (1.0 μg/ml) at the anaesthetist's discretion. Surgery was performed as a strictly monitor-based, non-rib-spreading procedure. After surgery, patients went to the post-anaesthesia care unit (PACU), where the ward nurse replaced the perioperatively used standard epidural medication with a blinded project infusion comprising either active medication or saline. When fully awake, patients were transferred to the surgical ward, continuing epidural infusion and baseline analgesics until the trial concluded. Post-operative pain assessment and treatment were ensured under the department’s usual standard of care, using numeric rating scale scores with on-demand intravenous "rescue" morphine available to all patients. The trial ended after blood sampling in the morning on the first post-operative day. At that time, post-operative epidural analgesia was changed from study medication to our normally used medication and continued until the concurrent removal of the chest drain and epidural catheter as per our department’s standard of care. After surgery, we performed follow-ups using the electronic patient chart, registering the cancer stage two weeks after surgery and post-operative survival two years after discharge.

Blood sampling, storage and analysis

We obtained venous blood samples at four time points. A baseline sample (T0) was obtained at least one day before surgery. A second sample was obtained when the surgeon stapled the pulmonary artery during surgery (T1). The third blood sample was drawn immediately upon arrival at the PACU (T2), and the final sample was obtained in the surgical ward on the first post-operative morning (T3). Blood sample storage, processing and cytokine measurements were done as described previously by Nielsen et al. [6] using electrochemiluminescence V-PLEX Custom Human Biomarker immunoassays (Mesoscale, Rockville, USA).

Endpoints

The primary outcome was changes in plasma levels of NK cells. The secondary outcomes were changes in NK cell activity, changes in plasma levels of interleukin (IL)-6, IL-10, IL-12, and interferon (IFN)-γ, the registered cancer stage and patient survival during the two-year follow-up period after surgery.

Statistical analyses

Studies on similar populations using comparable outcomes have found significant differences in the inflammatory parameters measured using sample sizes ranging from 20 to 57 patients [7, 8]. Using a pragmatic approach, we based our sample size on these studies and, placing ourselves in the upper range, decided on a minimum sample size of 28 patients in each group, a total of 56.

The participants' baseline demographics and trial data are presented as frequencies and percentages or as means and standard deviations (SDs). The cytokine level measurements were analysed using linear mixed models with and without the duration of the surgical procedure and the cancer stage as covariates, both including and excluding interaction effects between treatment groups and time. We evaluated the normality assumptions visually using the models' residuals. We used a log-rank test to examine whether a significant survival difference existed between the two treatment groups. All statistical analyses were done without imputation using Stata/BE 17.0 (StataCorp LLC, College Station, Texas, USA) software, and two-sided p-values below 0.05 were considered statistically significant.

Trial registration: NCT02359175 (ClinicalTrials.gov).

RESULTS

Patient flow

Patients were consecutively included from 17 January 2017 to 11 January 2019. The last patient was discharged from the hospital on 24 January 2019, thus concluding the two-year follow-up on 24 January 2021. As shown in Figure 1, we screened 253 scheduled VATS lobectomy patients and excluded 187 patients following the predefined criteria, leaving 66 for trial randomisation. Ten patients were excluded after inclusion, leaving 56 patients for the final analyses. Among these, 49 were alive at the two-year follow-up. The baseline patient characteristics and trial data are shown in Table 1. The patients' demographics and baseline data were comparable in the randomised groups, except for the American Society of Anesthesiologists (ASA) classifications and the Lung Cancer Stage Grouping, which were somewhat skewed. We examined experienced pain in the two treatment groups using the same definitions as in a previous publication. We found no statistically significant difference in experienced pain, either at rest (p > 0.05) or during activity (p > 0.06) [5]. The measured plasma levels of IL-6, IL-10, IL-12 and IFN-γ are aggregated in Figure 2. Summary statistics and levels of missing values are shown in the Supplementary File. The regression analyses of plasma cytokine levels are presented in Table 2. None of the examined cytokines exhibited significant differences in plasma concentrations between the two treatment groups when adjusted for time of blood sampling, p ≥ 0.204. However, cytokine levels varied significantly over time for all the examined interleukins. Three patients in the TE Group and four in the OM Group died during the two-year follow-up after surgery. The log-rank test showed no significant difference in survival between the two treatment groups (p = 0.5). The trial concluded with no patients experiencing serious adverse events or reactions.

DISCUSSION

This study found no difference in post-operative immunological markers or mortality in patients having VATS, regardless of whether the post-operative analgesic regimen comprised central neural blockade or oral morphine. IL-6 and IL-10 followed the expected course with a significant rise during and immediately after surgery. This was followed by a reduction towards the preoperative levels of the pro-inflammatory IL-6 and a post-operative increase in the anti-inflammatory IL-10 [9]. Plasma levels of IL-12 increased after surgery, reaching statistically significant elevations at T2. These findings run contrary to most previous results showing IL-12 depression after major vascular surgery [10], cardiac surgery [11] and orthopaedic surgery [10]. The post-operative reduction in the pro-inflammatory IFN-γ observed in our study aligns with what has been reported in other studies [12], including thoracic surgery [11].

We observed no between-group differences in any of the measured cytokines, showing no or only a minimal effect of epidural analgesia on post-operative cytokine levels in our setup. Previous studies have indicated an attenuating effect on post-operative immunosuppression when using epidural analgesia in both open and videoscopic surgery [13-15]. A 2017 trial on VATS lobectomies found an attenuating effect on the post-operative levels of IL-6 and INF-γ [16]. More recently, a study by Okuda and colleagues showed a significant between-group difference in IL-6 and IL-10 levels in the lung epithelial lining fluid but no differences in plasma levels [17]. We were unable to reproduce these findings in our study. However, most surgical procedures investigated in the studies were considerably more extensive than the thoracoscopic approach used in our trial, and the extent of surgery has been shown to influence the release of both pro-and anti-inflammatory cytokines [18, 19]. This may indicate that the extent of immune system activation in minimally invasive surgery is insufficient for epidural analgesia to exert any statistically significant effects on plasma cytokine levels. However, previous studies have examined the effects of epidural analgesia versus no epidural analgesia perioperatively, whereas our study used similar epidural analgesia in both groups perioperatively, changing to two different regimes post-operatively. Our setup might not allow epidural analgesia enough time to exert any attenuating effect on the immune response. Studies using a setup with a different perioperative epidural analgesia are needed to comprehensively examine any attenuating immune response attributable to epidural analgesia in VATS.

We found no difference in survival between the two groups using the log-rank test (p = 0.5). Evidence preceding our study has indicated a potential positive effect of epidural analgesia on post-operative mortality (odds ratio = 0.60) [3]. No studies on VATS existed at the time. Subsequently, a 2021 RCT on 400 VATS procedures in lung cancer patients showed no difference in post-operative survival using epidural analgesia compared with general anaesthesia alone [20]. Our study aligns with these results but is potentially underpowered to determine this with only seven events recorded during the two-year follow-up.

Our study has limitations that need to be addressed. First, our planned primary outcome measure was changes in plasma levels of NK cells after surgery. Because of technical difficulties in the laboratory, NK cell count and activity analyses were not performed. This might leave the study underpowered to detect differences in cytokine levels and post-operative survival. Second, the observed skewness in the baseline cancer stage and ASA classification might impact results. We speculate that a "healthier" OM Group, as indicated by the ASA classification, might reduce any ameliorating effect of epidural analgesia on post-operative immune suppression after cancer surgery. Conversely, the higher cancer staging in this group may suggest a more severe immunosuppression, which could potentially lead to an additional beneficial effect in this group due to baseline skewness. This, however, remains pure speculation. In the regression analysis, the duration of the surgical procedure and the patient's cancer stage did not add significantly to the regression model and were omitted from the final analysis. Likewise, the interaction effect between the treatment group and time was insignificant for all cytokines and was therefore omitted.

CONCLUSION

No significant differences were seen in the plasma cytokine levels or the two-year post-operative survival in VATS lobectomy when comparing epidural analgesia to enteral opioids as post-operative analgesics.