1. INTRODUCTION

Lung cancer is the most commonly diagnosed malignancy and the leading cause of cancer death worldwide, accounting for approximately 1.8 million deaths in 2020.1 Although surgical resection is a potentially curative treatment for nonmetastatic lung cancer, about 15.0% to 38.5% of patients develop locoregional recurrence 5 years after radical resection, which significantly worsens survival and quality of life.2

Retrospective studies have demonstrated an association between regional analgesia using local anesthetics with lower recurrence and better survival after tumor resection in certain types of cancer3,4; however, these findings have not been replicated in recent randomized controlled trials.5,6 Reduced stress responses and the use of opioids have been proposed to account for the benefits of epidural anesthesia and analgesia on cancer outcomes.3,4 Furthermore, preclinical studies have reported that local anesthetics could exert both tumoricidal and tumor-promoting effects on lung cancer cells depending on the concentration and type of local anesthetic, and the experimental model used.7–10 Various mechanisms have been proposed for the tumor-modifying effect of local anesthetics, including the upregulation of hypoxia-inducible factors,7 downregulation of inflammatory Src-signaling,8 inhibition of Akt and focal adhesion kinase,9 and blockade of voltage-gated sodium channels10 in cancer cells. A recent randomized controlled trial demonstrated that an intravenous bolus of lidocaine 1.0 mg·kg–1 over 10 minutes and continuous infusion at a rate of 1.0 mg·kg–1·h–1 throughout surgery led to a reduction in serum levels of interleukin-17 and cortisol, and reduction in pain intensity after video-assisted thoracic surgery for nonsmall-cell lung cancer.11 Similarly, another trial showed that an intravenous bolus of lidocaine 1% (1.5 mg·kg–1), continuous infusion of lidocaine 2 mg·kg–1·h–1 during surgery, and 1 mg·kg–1·h–1 for 24 hours postoperatively reduced the expressions of neutrophil extracellular trapping and matrix metalloproteinase 3 after surgical resection of primary breast tumors.12 In addition, a retrospective study reported that intravenous loading of lidocaine (1.5 mg·kg–1) during induction of anesthesia and continuous intraoperative infusion of 2 mg·kg–1·h–1 were associated with longer overall survival in patients following pancreatectomy for pancreatic cancer.13 However, few studies have investigated the relationship between local anesthetic dose and oncological outcomes after curative resection of lung cancer.

We performed this retrospective cohort study to examine the association between local anesthetic dose and lung cancer outcomes at a medical center in northern Taiwan. Based on the antimetastatic effects of local anesthetics reported in previous studies,3,4,8-13 we hypothesized that patients using more epidural bupivacaine would have a lower risk of cancer recurrence and all-cause mortality after surgical resection of nonsmall-cell lung cancer.

2. METHODS 2.1. Clinical setting and patient selection

This study was approved by the Research Ethics Committee of Taipei Veterans General Hospital, Taiwan (IRB-TPEVGH No. 2015-11-010CC). We included 2,581 consecutive patients who underwent lung resection surgery at Taipei Veterans General Hospital from 2005 to 2015 based on electronic medical records. The exclusion criteria were patients with benign lesions, metastatic lung cancer, small-cell lung cancer, distant metastasis diagnosed at the time of surgery, missing critical data, follow-up time <30 days, who did not use epidural patient-controlled analgesia, and those who received an epidural bupivacaine dose < 100 mg. After the exclusion process, a total of 464 patients who underwent surgical resection of primary stage IA to IIIB nonsmall-cell lung cancer were selected for further analysis (Fig. 1).

Fig. 1:

Flow diagram of patient selection.

2.2. Protocol of epidural analgesia

Epidural analgesia is commonly used to relieve postoperative pain after thoracic surgery at our hospital. The contraindications for epidurals are international normalized ratio >1.3, platelet count <100,000 μL–1, severe spine deformity, severe sepsis, and patient refusal, as described in our previous articles.14,15 For thoracic surgery, epidural catheters are typically placed between T6 and T8, threaded upwards 5 to 7 cm into the epidural space, and tested 1 day before surgery. During surgery, an epidural bolus of 5 to 10 mL lidocaine 1% is routinely administered before surgical incision, followed by a continuous infusion of bupivacaine 0.25% or 0.5% at a rate of 5 to 10 mL·h–1. Epidural analgesia with a patient-controlled pump (Gemstar Yellow, Hospira, Illinois, USA) is then given using bupivacaine 0.1% and fentanyl 1 μg·mL–1 with a demand dose of 2 to 2.5 mL and a basal infusion rate of 5 to 6 mL·h–1 for 2 to 3 days after surgery. Data of epidural bupivacaine consumption for each patient are recorded by the infusion pump machines and were retrieved for this study.

2.3. Clinical and pathological variables

Prognostic factors for cancer recurrence and all-cause mortality were collected from the electronic medical database of the hospital. The clinical variables were American Society of Anesthesiologists physical status, Eastern Cooperative Oncology Group grade,16 chronic obstructive pulmonary disease,17 preoperative forced vital capacity and forced expiratory volume in one second,18 pretreatment concentration of carcinoembryonic antigen (CEA),19 extent of surgical resection, use of the thoracoscopy technique, intraoperative blood loss, use of blood transfusions,20-22 and duration of anesthesia. Acute pain intensity on postoperative days 0 to 2 was measured using a self-reported 11-point numeric rating scale, with response options from “no pain” to “the worst pain.” The pathological variables were tumor differentiation, microscopic necrosis, lymphovascular invasion, perineural invasion, and cancer stage based on the American Joint Committee on Cancer staging system, 7th edition.23-25 Adjuvant cisplatin and carboplatin-based chemotherapy or radiation therapy were defined as the treatment within 90 days of surgery.

2.4. Outcome measurements

The primary outcome was disease-free survival, defined as the interval from surgery to the first cancer recurrence. Recurrence was defined as the presence of new intra- or extrapulmonary tumor nodules identified by spiral computed tomography, magnetic resonance imaging, bone scintigraphy, or positron emission tomography. The secondary outcome was overall survival, defined as the interval between surgery and death from any cause. The date of death was determined according to medical records and death certificates. The patients were followed until May 31, 2017. One author (Y.-H.T.) conducted a blinded assessment of the study outcomes.

2.5. Statistical analysis

The Shapiro-Wilk test and Kolmogorov-Smirnov test were used to test normality of the data. Skewed covariates were logarithmically transformed to conform to normality, including pretreatment CEA level, intraoperative blood loss, and duration of anesthesia. The associations between bupivacaine dose with disease-free survival and overall survival were assessed using proportional hazards regression models. Variables significantly associated with disease-free survival or overall survival in univariate analysis were incorporated into multivariate models to adjust for potential confounders. For sensitivity analysis, a stepwise forward model selection procedure was used with entry and removal significance criteria of 0.1 and 0.05, respectively, to determine the independent factors for disease-free survival and overall survival from the significant factors in univariate analysis. Finally, multivariate linear regression analysis was conducted to determine the factors associated with epidural bupivacaine dose. A two-sided significance level of 0.05 was used to define statistically significant differences. All statistical analyses were conducted using Statistics Analysis System (SAS) software, version 9.4 (SAS Institute Inc., Cary, NC, USA).

3. RESULTS

Among the 464 included patients, the mean dose of epidural bupivacaine was 352 mg (±standard deviation 74 mg). The minimum and maximum bupivacaine doses were 110 and 583 mg, respectively, and the interquartile range was from 311 to 402 mg. Tables 1 and 2 show the demographic, clinical and pathological characteristics of the included patients.

Table 1 - Demographic and clinical characteristics of the included patients Patient characteristics All patients (n = 464) Bupivacaine dose, mg 352 ± 74 Age, y 64 ± 11 Sex, male 213 (45.9%) Body mass index, kg·m –2 24.0 ± 3.4 ASA class ≥ 3 103 (22.2%) ECOG grade ≥ 1 107 (23.1%) Comorbidities  COPD 28 (6.0%)  Diabetes mellitus 77 (16.6%)  Coronary artery disease 45 (9.7%)  Heart failure 26 (5.6%)  Stroke 18 (3.9%)  Chronic kidney disease 38 (8.2%) Pulmonary function test  FVC, L 2.89 ± 0.76 FVC, % predicted 87.3 ± 14.8 FEV1, L 2.25 ± 0.61 FEV1, % predicted 86.5 ± 15.7 Pretreatment CEA, μg·L –1 a 2.3 (1.5 – 3.9) Preoperative laboratory tests Hemoglobin concentration, g·dL-1 13.1 ± 1.4 Platelet count, 103·μL–1 217 ± 71 International normalized ratio 1.01 ± 0.05 Surgical and anesthetic variables Type of surgery Sublobar resection 114 (24.6%) Lobectomy 343 (73.9%) Bilobectomy or pneumonectomy 7 (1.5%) Thoracoscopic surgery 441 (95.0%) Radical lymph node dissection 360 (77.6%) Intraoperative blood loss, mLb 50 (30 – 150) Perioperative blood transfusion 36 (7.8%) Anesthesia duration, minb 300 (240 – 345) Daily maximum NRS pain score Postoperative day 0 2.2 ± 1.8 Postoperative day 1 2.6 ± 1.9 Postoperative day 2 2.4 ± 1.4 Postoperative day 0-2 average 2.4 ± 1.3 Postoperative LOS, d 8 (7 – 10) Year of operation  2005–2010 56 (12.1%)  2011–2015 408 (87.9%) Preoperative C/T ± R/T 13 (2.8%) Postoperative C/T 195 (42.0%) Postoperative R/T 19 (4.1%)

Values were mean ± standard deviation, counts (percent), or median (interquartile range).

ASA = American Society of Anesthesiologists; CEA = carcinoembryonic antigen; COPD = chronic obstructive pulmonary disease; C/T = chemotherapy; ECOG = Eastern Cooperative Oncology Group; FEV1 = forced expiratory volume in one second; FVC = forced vital capacity; LOS = length of hospital stay; NRS = numeric rating scale; R/T = radiotherapy.

aOn base-10 logarithmic scale.

bOn base-2 logarithmic scale.

Table 2 - Cancer stages and pathological characteristics Stage and pathology All Patients (n = 464) AJCC stage  Stage I 373 (80.4%)   IA 175 (37.7%)   IB 198 (42.7%)  Stage II 44 (9.5%)   IIA 27 (5.8%)   IIB 17 (3.7%)  Stage III 47 (10.1%)   IIIA 45 (9.7%)   IIIB 2 (0.4%) Pathological characteristics  Subtype  Adenocarcinoma 405 (87.3%)  SCC 37 (8.0%)   Other 22 (4.7%)  Tumor differentiation   Good 71 (15.3%)   Moderate 259 (55.8%)   Poor 134 (28.9%)  Microscopic necrosis 63 (13.6%)  Lymphocytic infiltration 16 (3.5%)  Lymphovascular invasion 117 (25.2%)  Perineural infiltration 10 (2.2%)

Values were counts (percent).

AJCC = American Joint Committee on Cancer; SCC = squamous cell carcinoma.

3.1. Bupivacaine dose and disease-free survival

The results of univariate analysis for disease-free survival are shown in Supplementary Table 1, https://links.lww.com/JCMA/A157. In multivariate analysis, the association between bupivacaine dose and cancer recurrence was not significant (adjusted hazard ratio: 1.000, 95% confidence interval: 0.997–1.002, p = 0.771), similar to the results of forward model selection (Table 3 and Supplementary Table 2, https://links.lww.com/JCMA/A157). Independent factors associated with cancer recurrence included cancer stage, tumor differentiation, lymphovascular invasion, heart failure, postoperative chemotherapy, and postoperative radiotherapy.

Table 3 - Multivariable analyses for cancer recurrence and all-cause mortality Cancer recurrence HR (95% CI) p All-cause mortality HR (95% CI) p Bupivacaine dose, mg 1.000 (0.997–1.002) 0.771 Bupivacaine dose, mg 1.008 (1.001–1.016) 0.029 COPD 0.821 (0.277–2.432) 0.722 Age, y 1.062 (0.990–1.139) 0.092 Diabetes mellitus 1.258 (0.761–2.080) 0.372 Sex, male 1.098 (0.373–3.237) 0.865 Heart failure 2.178 (1.074–4.416) 0.031 ASA class ≥ 3 1.177 (0.424–3.270) 0.755 FVC, % predicted 0.498 (0.039–6.411) 0.593 ECOG grade ≥ 1 1.318 (0.407–4.271) 0.646 FEV1, % predicted 0.405 (0.031–5.365) 0.493 Stroke 1.171 (0.298–4.598) 0.821 Pretreatment CEA, μg·L–1a 1.037 (0.639–1.682) 0.885 Pretreatment CEA, μg·L –1 a 1.949 (0.726–5.232) 0.185 Hemoglobin level, g·dL –1 0.958 (0.820–1.119) 0.587 Type of surgery 0.069 Platelet count, μL –1 1.000 (1.000–1.000) 0.550  Sublobar resection (reference) Type of surgery 0.440  Lobectomy 0.255 (0.056–1.162) 0.077  Sublobar resection (reference)  Bilobectomy or pneumonectomy 3.567 (0.097–130.53) 0.489  Lobectomy 0.533 (0.178–1.598) 0.261 Intraoperative blood loss, mL b 1.414 (0.823–2.429) 0.210  Bilobectomy or pneumonectomy 0.343 (0.056–2.087) 0.245 Perioperative blood transfusion 1.442 (0.250–8.307) 0.682 Radical lymph node dissection 1.743 (0.585–5.192) 0.319 Anesthesia duration, min b 0.664 (0.120–3.690) 0.640 Anesthesia duration, min b 1.490 (0.847–2.624) 0.167 Postoperative LOS, d 1.089 (1.012–1.172) 0.023 Preoperative C/T ± R/T 0.556 (0.213–1.447) 0.229 Preoperative C/T ± R/T 0.602 (0.022–16.787) 0.765 Postoperative C/T 1.837 (1.097–3.077) 0.021 Postoperative C/T 2.282 (0.559–9.318) 0.250 Postoperative R/T 2.394 (1.178–4.864) 0.016 Postoperative R/T 1.799 (0.384–8.428) 0.456 Cancer stage <0.001 Cancer stage 0.052  II vs. I 2.459 (1.352–4.474) 0.003  II vs. I 4.975 (1.058–23.386) 0.042  III vs. I 3.280 (1.815–5.926) <0.001  III vs. I 6.420 (1.303–31.635) 0.022 Subtype 0.116 Tumor differentiation 0.330  SCC vs. adenocarcinoma 0.652 (0.250–1.699) 0.381  Moderate vs. good . (0.000–.) 0.993  Other vs. adenocarcinoma 0.429 (0.190–0.971) 0.042  Poor vs. good . (0.000–.) 0.992 Tumor differentiation 0.010 Microscopic necrosis 1.663 (0.588–4.702) 0.338  Moderate vs. good 4.032 (0.942–17.255) 0.060 Lymphocytic infiltration 2.083 (0.490–8.850) 0.320  Poor vs. good 6.836 (1.532–30.498) 0.012 Lymphovascular invasion 0.708 (0.214–2.335) 0.570 Microscopic necrosis 1.053 (0.620–1.788) 0.849 Lymphovascular invasion 2.058 (1.250–3.386) 0.005 Perineural infiltration 1.267 (0.491–3.268) 0.625

ASA = American Society of Anesthesiologists; CEA = carcinoembryonic antigen; CI = confidence interval; COPD = chronic obstructive pulmonary disease; C/T = chemotherapy; ECOG = Eastern Cooperative Oncology Group; FEV1 = forced expiratory volume in one second; FVC = forced vital capacity; HR = hazard ratio; LOS = length of hospital stay; R/T = radiotherapy; SCC = squamous cell carcinoma.

aOn base-10 logarithmic scale.

bOn base-2 logarithmic scale.

3.2. Bupivacaine dose and overall survival

The results of univariate analysis for overall survival are shown in Supplementary Table 1, https://links.lww.com/JCMA/A157. Multivariate analysis demonstrated a significant association between bupivacaine dose and all-cause mortality (adjusted hazard ratio: 1.008, 95% confidence interval: 1.001–1.016, p = 0.029) (Table 3). To further confirm this finding, the forward model selection procedure was used, which showed similar results (adjusted hazard ratio: 1.007, 95% confidence interval: 1.000–1.014, p = 0.037) (Supplementary Table 2, https://links.lww.com/JCMA/A157). Other independent prognostic factors for overall survival were cancer stage and postoperative length of hospital stay.

3.3. Factors associated with bupivacaine dose

Five factors were found to be independently associated with postoperative bupivacaine dose, namely age, sex, body mass index, mean daily maximum numeric rating scale pain score, and pathological perineural infiltration (Table 4).

Table 4 - Influential factors of bupivacaine dose Influential factor Beta 95% CI p Age –1.514 –2.271 to –0.757 <0.001 Sex, male vs. female 36.241 18.999 to 53.483 <0.001 Body mass index, kg·m−1 3.707 1.117 to 6.297 0.005 Mean daily maximum NRS pain score 13.567 7.465 to 19.669 <0.001 Perineural infiltration 59.041 3.136 to 114.946 0.039

CI = confidence interval; NRS = numeric rating scale.

4. DISCUSSION

Our results showed a dose-dependent association between postoperative bupivacaine dose and mortality risk after surgical resection of nonsmall-cell lung cancer, in contrast to the hypothetical anticancer effects of local anesthetics. In this study, we carefully recruited highly homogeneous patients with regards to disease characteristics and clinical setting. In addition, we adjusted for a comprehensive list of clinical and pathological variables in our analysis to minimize potential confounding effects. Our findings may provide important clinical implications for perioperative pain management in cancer surgery.

Although epidural blockade has been associated with better oncological outcomes in colorectal cancer, prostate cancer, and breast cancer,3,4 this beneficial effect has not been found in patients undergoing curative resection of nonsmall-cell lung cancer.14,26 Importantly, the relationship between the epidural dose of local anesthetics and cancer outcomes has not been examined in previous studies.14,26 In the present study, we observed a higher risk of all-cause mortality but not cancer recurrence in the patients who used more epidural bupivacaine for postoperative analgesia. There are several possible explanations for this finding. First, we controlled for many important covariates in our analysis, which may have affected the influence of bupivacaine dose on recurrence. Second, although we considered a variety of potential confounders in the analysis, competing causes of death were still possible. The higher use of bupivacaine may reflect more extensive surgical resection and complicated postoperative course rather than the dose of bupivacaine itself. Alternatively, the patients receiving higher dosages of bupivacaine may have had a higher rate and longer duration of hypotension perioperatively, which is a known predictor of mortality.27

Reduced stress responses and preserved immunity have been proposed to account for the benefits of regional anesthesia on cancer outcomes.3,4 Although the mechanisms underlying the protective effect of regional anesthesia remain unclear,3,4 one possibility is the anticancer effect of local anesthetics per se. Experimental studies have shown that local anesthetics may exert mixed effects on the behavior of tumor cells.7-12 Piegeler and coworkers reported that ropivacaine and lidocaine treatment downregulated the activity of biomarkers of tumor growth and metastasis in lung cancer cells.8,9 In addition, Hou et al reported that an intraoperative infusion of intravenous lidocaine attenuated systemic inflammation and alleviated acute pain after surgical resection of early-stage nonsmall-cell lung cancer, but that the rate of cancer recurrence was not affected.11 Moreover, Galoș and colleagues reported that an intravenous infusion of lidocaine decreased the expressions of angiogenesis and neutrophil extracellular trapping biomarkers in breast cancer.12 Conversely, Chan and coworkers reported that levobupivacaine treatment upregulated the hypoxia-inducible factor-2α gene, triggered epithelial-to-mesenchymal transition, and promoted the dissemination of lung cancer cells.