Impact of opioid-free general anesthesia with locoregional blocks during breast cancer surgery on systemic inflammatory response: results from a randomized controlled trial

Abstract

Background

For breast cancer patients with localized disease, surgery remains a cornerstone of treatment. Anesthesia during surgery is essential for pain management and influences patient recovery. Opioid-free general anesthesia with locoregional blocks (OfGA), as an alternative to traditional opioid-based general anesthesia (GA), may improve patient outcomes.

Methods

We set up a randomized trial to compare the impact of different anesthetic approaches on post surgical inflammation in breast cancer patients undergoing quadrantectomy. Patients were randomized to receive GA (n=34) or OfGA (n=34). Blood samples were collected pre-surgery and 1 hour and 24 hours post-surgery. As primary endpoint, neutrophil-to-lymphocyte ratio (NLR) was calculated. As secondary endpoints we collected data on side effects, pain management, and circulating inflammatory cytokines.

Results

Patients of the GA arm displayed significantly increased NLR, both at 1 (median change +0.13, IQR 0.01 to 0.62, p<0.0001) and 24 (median change +0.64, IQR 0.28 to 1.03, p<0.0001) hours post-surgery, compared with baseline. On the contrary, the OfGA arm displayed stable or minimally increased NLR (median change -0.08, IQR -0.21 to 0.17, p=ns, at 1 hour; +0.11, IQR -0.11 to 0.36, p=ns, at 24 hours post-surgery). Circulating inflammatory cytokines aligned with NLR data, showing significant increases in key inflammatory factors in samples from the GA arm compared with the OfGA arm. Pain management was similar in both arms, but OfGA was better tolerated and elicited fewer side effects.

Conclusions

In breast cancer surgery, opioid-free general anesthesia combined with locoregional blocks was associated with a reduced postoperative inflammatory response compared with conventional opioid-based general anesthesia, without compromising pain control. Further research is needed to determine whether these biological effects translate into meaningful clinical outcomes.

Trial registration number

NCT04172220.

WHAT IS ALREADY KNOWN ON THIS TOPIC

• Anesthesia is crucial for surgery, pain control, and patient recovery. Efforts are increasingly focused on opioid-free approaches to reduce associated adverse effects.

WHAT THIS STUDY ADDS

• This randomized trial in breast cancer patients undergoing quadrantectomy shows that a locoregional approach with opioid-free general anesthesia provides optimal pain control and better tolerance and reduces post-surgical inflammation compared with opioid-based general anesthesia.

HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY

• This study contributes to the growing body of evidence supporting opioid-sparing anesthetic techniques. While it shows that opioid-free general anesthesia with locoregional blocks may reduce short-term markers of inflammation, further investigation is needed to clarify whether these changes impact long-term clinical outcomes or should influence routine anesthetic practice in breast cancer surgery.

Introduction

Breast cancer (BC) is the most common malignancy in women, and surgical resection is a cornerstone of treatment, playing a pivotal role in determining prognosis and long-term outcomes.¹ In this contest, anesthesia management is essential for the surgical procedure and it can also influence postoperative pain control and patient recovery.

Breast surgery is commonly performed under general anesthesia (GA) alone or in combination with locoregional blocks. Traditional locoregional techniques, such as thoracic epidural and paravertebral blocks, are highly effective in managing intraoperative and postoperative pain and reducing systemic analgesic requirements, but can be limited by specific patient conditions and potential complications.^{2 3} Chest wall blocks have emerged as a promising alternative, offering effective perioperative analgesia with a lower risk profile.^{4 5}

Recently, opioid-free general anesthesia combined with locoregional blocks (OfGA) has emerged as an interesting alternative to traditional opioid-based GA.⁶ The main objective of OfGA is to mitigate the adverse effects commonly associated with opioid use, including nausea, vomiting, respiratory depression, and immunosuppression. This approach may be especially valuable in oncological surgeries, where integrating regional anesthesia with opioid-sparing strategies may have the potential to improve long-term outcomes.

The interplay between cancer, inflammation, and immune response, particularly in the perioperative period, is a critical topic.^{7 8} Preclinical studies suggest that opioids can impair natural killer cell function, promote cancer recurrence by influencing angiogenesis, and modulate tumor cell signaling pathways.^{9 10} In this context, several inflammatory biomarkers have been investigated to improve prognostic accuracy. One such marker, the neutrophil-to-lymphocyte ratio (NLR), easily derived from a standard complete blood count, has emerged as a reliable indicator of systemic inflammation and immune competence. Studies suggest that elevated NLR may be not only a marker of inflammation but also a prognostic indicator in cancer patients.^11–14 Accordingly, higher NLR values have been associated with more aggressive disease characteristics, such as advanced tumor stages and greater metastatic burden, which correlate with higher risks of complications and poorer overall survival.¹⁵ A meta-analysis of 100 studies demonstrated that elevated pre-treatment NLR values are strongly associated with unfavorable prognosis and reduced overall survival across a range of cancer types and stages.¹⁶ The link between a high NLR and worse oncological outcome is often attributed to immune system alterations.^{10 17–24} Opioid use, particularly during GA, has been linked to increases in NLR, suggesting a temporary phase of postoperative immunosuppression. Moreover, the expression of opioid receptors on tumor cells raises the possibility of opioids directly influencing tumor behavior,^{25 26} further underscoring the complex interplay between inflammation, immune modulation, and cancer progression, in this context.

Many clinical studies have proposed that regional anesthesia contributes to better outcomes in cancer patients, by mitigating the surgical stress response, reducing the need for volatile anesthetics and opioids, thus modulating adrenergic and inflammatory pathways, while augmenting immune surveillance.^27–29 The actual translation of these beneficial effects on patients’ survival is still controversial.^30–32 Despite that the debate on the impact of opioids on cancer prognosis continues, research on OfGA remains a dynamic and evolving field. Its potential for minimizing opioid-related adverse effects while addressing the growing concerns surrounding opioid use and abuse represents a promising development in perioperative oncology care.

In this study, we present findings from a randomized clinical trial involving a homogeneous and carefully selected cohort of BC patients undergoing quadrantectomy. We compared the effects of traditional opioid-based GA with those of a locoregional approach combined with opioid-free general anesthesia (OfGA). Systemic inflammatory responses were measured through the assessment of NLR and circulating cytokine levels, at multiple time points before and after surgery, disclosing significant advantages in the OfGA arm.

Methods

Study design

Single-center open-label randomized controlled trial with two parallel arms (1:1 ratio) conducted by the Breast Surgery Unit of the National Cancer Institute, Centro di Riferimento Oncologico (CRO) di Aviano, IRCCS, PN, Italy. The first patient was enrolled on November 27, 2019. When already the first 23 patients were enrolled, an amendment to introduce an additional secondary objective was approved by the Regional Ethics Committee (CEUR, Prot. N. GEN-GEN-2020-0011494-A, dated July 1, 2020). From that moment, 45 more patients were enrolled, reaching a total of 68 patients. The last patient was enrolled on September 28, 2021. The protocol and trial conduct complied with the Declaration of Helsinki, International Council for Harmonization E6 Guideline for Good Clinical Practice and European regulations.

Objectives

The primary objective of the study was to determine whether the two groups, treated with different anesthetic approaches, differed in the variation of the NLR, measured between the postoperative period (24 hours after surgery) and the preoperative baseline. The two groups were as follows: locoregional anesthesia (interpectoral and SAP blocks) combined with opioid-free general anesthesia (altogether referred to as “OfGA” arm; n=34 patients); conventional general anesthesia with opioids (referred to as “GA” arm; n=34 patients).

The secondary objectives were as follows: (1) evaluating the trend of NLR values over time in both groups, considering preoperative values, measurements taken 1 hour after surgery, and values recorded 24 hours postoperatively; (2) comparing the extent of postoperative pain symptoms between the two arms, with the assessment of opioid consumption; (3) comparing the incidence of side effects, such as nausea and vomiting, using the validated Numeric Rating Scale (NRS); (4) exploring whether variations in NLR were accompanied by corresponding changes in circulating cytokine levels during the postoperative period (1 hour and 24 hours after surgery), relative to preoperative baseline levels. This last secondary objective was introduced after the trial was already started (see above for amendment details) and therefore involved only the last 45 patients (24 in the GA arm and 21 in the OfGA arm).

Study population

Eligible patients were adult women with a first diagnosis of BC, scheduled for external quadrantectomy and axillary surgery (sentinel lymph node biopsy and possible axillary lymphadenectomy), not undergoing any kind of treatment involving opioids. Further inclusion criteria were as follows: ability to express adequate and specific informed consent; intact cognitive skills. Exclusion criteria were as follows: pregnancy at the time of the study; antineoplastic therapy or follow-up for another malignancy at the time of the study; ongoing therapy with opioids or other category of analgesic drugs at the time of the study; immune-rheumatological disease; documented allergy to local anesthetics or previous adverse reactions; documented history of difficult airways or any anesthetic problem during previous surgeries; predictable inadequacy to ensure correct compliance with the study for logistical or psychological reasons; refusal to participate in the clinical study and/or in subsequent blood sampling planned by the clinical protocol.

In all patients, we determined the variation of NLR, assessing the value of this index in the postoperative period compared with the baseline value (figure 1). Our hypothesis was that, in patients belonging to the OfGA group, a smaller increase in the NLR value is predictable in the postoperative period compared with the baseline preoperative value, in comparison with the patients belonging to the GA group.

Figure 1. Consolidated Standards of Reporting Trials (CONSORT) diagram of the study. A randomized controlled clinical trial was designed to assess the impact of opioid-free general anesthesia (OfGA) versus general anesthesia with opioids (GA) on post surgical inflammation in breast cancer (BC) patients undergoing quadrantectomy. For the primary outcome—variation in the neutrophil/lymphocyte ratio (NLR)—68 BC patients were enrolled, with 34 assigned to the GA arm and 34 to the OfGA arm. For the secondary outcome, evaluating variations in inflammatory cytokine levels, only the last 45 enrolled patients were included: 24 in the GA arm and 21 in the OfGA arm.

Intervention

Patients in both groups received 1–2 mg midazolam intravenous as premedication upon arrival in the operating room. In the OfGA group, locoregional anesthesia—interpectoral and serratus anterior muscle plane (SAP) blocks—was performed with an ultrasound-guided technique. For interpectoral block, 10 mL of local anesthetic (0.5% ropivacaine) was infiltrated between the fasciae of the pectoralis major and pectoralis minor muscles at the level of the third rib on the anterior axillary line, aiming to block the lateral and medial pectoral nerves. To block the SAP, we proceeded with the identification of the fifth rib on the middle axillary line and with the subsequent infusion of about 20 mL of local anesthetic (0.5% ropivacaine) between the fasciae of the latissimus dorsi muscle and serratus anterior muscle to block the thoracodorsal nerve, the long thoracic nerve, and the intercostal nerves T2 to T9. Chest wall blocks were performed maintaining the patient in the supine position with the arm abducted at 90°, with ultrasound guidance, a high-frequency linear probe (10–12 MHz), and a medial to lateral in-plane needle direction, using echogenic needles. After performing locoregional anesthesia, induction of general anesthesia was achieved with propofol 1.5–2 mg/kg intravenous. In the GA group, general anesthesia was induced with propofol 1.5–2 mg/kg intravenous, fentanyl 1 µg/kg intravenous (0.1 morphine milligram equivalents (MME) per kg). In both cases, to control the airway, a laryngeal mask was placed, with possible administration of rocuronium 0.6 mg/kg intravenous. Anesthesia was maintained with propofol according to Target-controlled infusion mode, modulated to maintain a Bispectral Index value between 40 and 60. The administration of fentanyl 50 µg (5 MME) as a bolus intravenous was contemplated in case of changes in arterial pressure and heart rate suggestive of painful stimulation (increase in mean arterial pressure or heart rate of at least 20% compared with preoperative values). Postoperative analgesia was ensured by the administration of paracetamol 1000 mg intravenous every 8 hours for the first 24 hours, ketorolac 30 mg intravenous in the event of poor pain control (NRS >4) administrable at most every 8 hours (max 90 mg/24 hours), and morphine 10–20 mg intravenous within 24 hours, in case of still uncontrolled pain.

Patients’ assignment to groups (randomization and allocation concealment) and enrollment

All patients who satisfied the inclusion and exclusion criteria and gave their informed consent to participate in the study were randomized into two arms: OfGA and GA.

A randomization list with randomly permuted blocks (seven blocks) was generated by an independent statistician using online software (www.randomization.com). Allocation concealment was granted by use of sealed, opaque, consecutively numbered envelopes containing assignation arm. For each eligible patient, the principal investigator opened the envelope with the lowest number and randomized the patient to the assigned treatment. Due to different procedures for the two anesthetic techniques, blinding was not possible.

Patients were enrolled on the day they underwent the preoperative anesthesiologic examination. All patients scheduled for external quadrantectomy and axillary surgery with sentinel lymph node biopsy and possible lymphadenectomy of the axilla were evaluated by the anesthesiologist for their eligibility to the study. Each patient was assigned a unique and anonymous identification code, which identified her in all data collection forms. Each patient enrolled in the study was evaluated preoperatively and followed up for the first 24 postoperative hours. To file the data concerning the development of chronic post-surgical pain, we used the Douleur Neuropathique 4 (DN4) questionnaire and patients were interviewed during the regular follow-up visits or reached by phone, at 3–6 and 12–24 months.

Calculation of neutrophil-to-lymphocyte ratio (NLR) and plasma preparation

Blood samples were collected before surgery, 1 hour after the end of surgery, and on the first postoperative day (24 hours after surgery). Complete blood count with leukocyte formula was performed. The NLR of all patients (n=68) was calculated based on the numbers of neutrophils and lymphocytes in the complete blood count and calculated by dividing absolute neutrophil count by absolute lymphocyte count. A fraction of the same blood drawing was used for the evaluation of the plasma levels of human cytokines linked to inflammation, on n=45 patients.

To obtain plasma, whole blood samples collected in EDTA tubes were processed within 2 hours from drawing and kept at 4°C until centrifugation. Whole blood sample was first centrifuged at 2000 rpm, and then supernatant was subjected to two more steps of centrifugation at 4000 rpm, in order to remove all cellular and subcellular contaminants. Plasma was then divided into aliquots and stored at −80°C, until analyzed in the Luminex assay.

Luminex assay

Circulating cytokines were analyzed in plasma collected from the last 45 patients enrolled in the study. We used the Luminex technology to simultaneously analyze 44 cytokines involved in the inflammation process, following the manufacturer’s protocol (Human XL Cytokine Luminex Performance Panel, RnD System, Biotechne) (full list in online supplemental table 1). Briefly, 50 μl of plasma was aliquoted in duplicates into the microplate wells, to obtain the binding of plasmatic analytes with specific antibodies pre-coated onto magnetic microparticles embedded with fluorophores. Samples, standards, and internal controls were then processed, plates were read using the Luminex Xponent analyzer, and data analyzed, following the manufacturer’s instructions.

Statistical considerations

Sample size

Sample size calculation was derived from pilot data collected prior to the design of the clinical trial, involving 10 patients for each group of anesthetic protocol. These preliminary findings provided an estimate of the expected magnitude of difference in NLR changes between groups, which informed our power analysis, leading to the assumption of a medium-to-large effect size (d=0.7). We determined that, to achieve a statistically significant difference between the two study arms in the mean intra-patient NLR values (primary objective), a statistical power (β) of 80%, and a two-sided significance level (α) of 5%, it was necessary to enroll 34 patients per arm. This calculation was based on the formula d = (mean A − mean B) / SD, where A and B represent the two study arms.

The analysis of the 44 inflammatory cytokines, a secondary objective of the study, was performed on the plasma from the last 45 enrolled patients (n=24 in the GA arm; n=21 in the OfGA arm).

Statistical analysis

The analysis is carried out with an intention-to-treat method. Data on patients’ characteristics are described with appropriate indexes of position (mean, median) and dispersion (SD, IQR) and as absolute and relative frequencies, according to the type of variables (table 1). For categorical or dichotomous variables, the Mann-Whitney U test or the Student t-test was used.

Table 1. Clinico-pathological characteristics of breast cancer patients enrolled in the study.

	GA	OfGA	P value
Number of patients	34	34
Age, years (median, IQR)	53 (46.5–69.0)	57 (50.8–72.5)	0.171
Menopausal status (n, %)
Pre	8 (24%)	8 (24%)	1.000
Post	26 (76%)	26 (76%)
Surgery duration, min (mean±SD)	92.4±24.8	85.7±19.9	0.229
Baseline NLR (median, IQR)	1.9 (1.6–2.4)	2.3 (1.8–3.0)	0.0228
Tumor histology (n, %)			0.528
Ductal	21 (62%)	25 (73%)
Lobular	9 (26%)	5 (15%)
Other (or not available)	3 (12%)	4 (12%)
Tumor subtype (n, %)			0.337
Luminal (ER+, HER2−)	28 (82%)	26 (76%)
HER2 (ER+/−, HER2+)	2 (6%)	6 (18%)
Triple-negative (ER−, HER2−)	4 (12%)	2 (6%)
Tumor grade (n, %)			1.000
G1	1 (3%)	2 (6%)
G2	22 (65%)	21 (62%)
G3	11 (32%)	11 (32%)
Tumor N (n, %)			0.235
0	25 (73%)	27 (79%)
1	5 (15%)	7 (21%)
2	3 (9%)	0 (0%)
3	1 (3%)	0 (0%)

GA, general anesthesia with opioids; NLR, neutrophil-to-lymphocyte ratio; OfGA, opioid-free general anesthesia with locoregional blocks.

Differences in quantitative variables (including principal outcome, NLR) between arms have been evaluated by Mann-Whitney U test, in the event of non-normality of the distribution. Normality assumption was tested both visually and with the Shapiro-Wilk test. Median quantile regression was carried out to take into account baseline abnormal distribution. Fold change over baseline NLR values was calculated to account for unbalance between arms in baseline values. Differences between time points were assessed with the Wilcoxon test for repeated measures.

A two-sided p value <0.05 was considered statistically significant.

Results

From November 2019 to September 2021, 68 patients have been enrolled in the trial and randomized into the two groups: 34 patients received interpectoral+SAP chest wall block and opioid-free GA (OfGA) and 34 received GA, with intraoperative fentanyl, as per protocol. Each patient adhered to their assigned treatment, with no crossover cases (figure 1).

All patients underwent breast external quadrantectomy with sentinel lymph node biopsy. A complete axillary lymphadenectomy was performed in one patient from each group and no mastectomies were necessary. The two groups were comparable in age (median age 57 in GA arm and 53 in OfGA arm) (table 1).

The preplanned primary outcome was to determine whether the change in NLR, measured from the preoperative baseline to 24 hours postoperatively, was influenced by the anesthetic technique. Since baseline NLR values did not meet the assumption of normality (p<0.001), we used non-parametric methods to compare median values. The absolute difference between median NLR at baseline and 24 hours post-surgery was 1.16 in GA (IQR 0.52 to 1.70) and 0.22 in OfGA (IQR −0.26 to 0.83), with a significant between-group difference (p<0.001). NLR increase remained significantly higher in the GA group compared with the OfGA group after adjusting for baseline values using a multivariate median regression model (adjusted coefficient for GA vs OfGA: 0.83, p=0.011). Direct comparison of median NLR values between groups at 24 hours was 0.53, with higher—though not statistically significant—values in the GA group (p=0.0503) (figure 2A).

Figure 2. The neutrophil-to-lymphocyte ratio (NLR) inflammation marker is lower in breast cancer patients treated with opioid-free general anesthesia (OfGA) compared to those treated with general anesthesia (GA). (A) Graph reports the NLR in liquid biopsies (LB) collected from patients at different time points (LB1, basal pre-surgery; LB2, 1 hour after surgery; LB3, 24 hours after surgery) after administration of GA or OfGA, as indicated. (B) Graph reports the NLR relative change at 1 (LB2) and 24 hours (LB3), normalized over baseline values (LB1), in GA and OfGA cohorts. In all box plots, boxes include the first and third quartiles, line indicates the median, and whiskers indicate the minimum and maximum data values. Outliers are reported at the extremities. Statistical significance was calculated with the Wilcoxon test (A) or Mann-Whitney U test (A, B), as more appropriate. In both graphs, straight lines indicate intra-arm comparisons; square brackets indicate inter-arm comparisons. *p<0.050, **p<0.01, ***p<0.001, and ****p<0.0001.

To address the unbalance of baseline NLR values (LB1 vs LB1, p=0.0228; see figure 2A and table 1), we reanalyzed data folding postoperative NLR values over their respective baseline ones. At 1 hour, the median change was +0.13 in GA (IQR 0.01 to 0.62, p<0.0001) versus –0.08 in OfGA (IQR −0.21 to 0.17, p=ns); at 24 hours, median change was +0.64 in GA (IQR 0.28 to 1.03, p<0.0001), compared with +0.11 in OfGA (IQR −0.11 to 0.36, p<0.05) (figure 2B). Intergroup comparisons of these adjusted NLR values confirmed significantly higher values in the GA arm, both at 1 hour (p<0.0002) and 24 hours (p<0.0001) post-surgery (figure 2B). Collectively, these findings suggest that depleting opioids coupled with regional blocks in OfGA attenuates the postoperative inflammatory response normally elicited by GA.

We next proceeded with the evaluation of pain symptoms; both groups had preoperative Numeric Rating Scale (NRS) value of 0. In the OfGA group, only three patients, one of whom underwent axillary lymphadenectomy, required a rescue dose of analgesic (ketorolac 30 mg intravenous) (table 2). In contrast, in the GA group, six patients (including the one who underwent axillary lymphadenectomy) needed rescue analgesia with ketorolac (30 mg intravenous); three of these six patients required an additional dose within 24 hours (30 mg intravenous). Furthermore, four GA patients required a 24-hour infusion of morphine 20 mg intravenous, due to a static NRS of 8, while none of the patients who received OfGa needed morphine (table 2). Regarding the side effects, the GA group exhibited a higher incidence of adverse symptoms: five patients experienced nausea, two had vomiting, three reported dizziness, one reported headache, and one reported drowsiness, while none of these symptoms were reported in the OfGA group (table 2). Finally, we evaluated the development of postoperative neuropathic pain, administering the DN4 questionnaire 3–6 and 12–24 months after surgery, during follow-up visits or interviewing patients by phone (table 2). At both time points, the number of patients reporting pain with a score above four was identical in the two groups, with five patients per arm (table 2).

Table 2. Survey on post-surgical side effects and acute/chronic pain management.

Patient ID #	Type of anesthesia	Hormone receptor status	Menopausal state	Surgery duration (minutes)	Nausea	Vomiting	Other side effects	Acute post-surgery pain		Need for pain-controlling drugs (24 hours post-surgery)	Neuropathic pain (DN4 questionnaire)
								1 hour	24 hours		3–6 months post-surgery	12–24 months post-surgery
1	OfGA	pos	post	60							0	0
2	OfGA	pos	pre	90							0	0
7	OfGA	pos	post	75							5	4
9	OfGA	pos	post	75							0	2
10	OfGA	pos	post	120							0	0
11	OfGA	pos	post	100							0	0
14	OfGA	pos	pre	145							0	0
15	OfGA	pos	post	105					X	Ketorolac tromethamine 30 mg	3	2
16	OfGA	pos	post	120							0	0
17	OfGA	pos	post	80							0	0
21	OfGA	pos	post	90							0	0
22	OfGA	pos	post	70							0	0
23	OfGA	neg	post	50							0	0
28	OfGA	pos	post	110							5	5
31	OfGA	pos	post	85							0	0
32	OfGA	pos	pre	70							5	0
35	OfGA	pos	post	75							0	0
36	OfGA	pos	post	60							0	0
37	OfGA	pos	post	70							0	0
39	OfGA	pos	post	100							0	0
41	OfGA	pos	post	90							0	0
42	OfGA	pos	post	110							0	0
43	OfGA	neg	pre	75							0	0
44	OfGA	neg	pre	90							6	6
47	OfGA	pos	post	75							0	0
51	OfGA	pos	post	65							0	0
53	OfGA	pos	post	75							0	0
58	OfGA	pos	post	85							0	0
59	OfGA	neg	post	80					X	Ketorolac tromethamine 30 mg	0	0
60	OfGA	pos	pre	95							0	0
61	OfGA	pos	pre	80							0	0
63	OfGA	pos	pre	90							6	6
64	OfGA	pos	post	65							NA	NA
66	OfGA	pos	post	90					X	Ketorolac tromethamine 30 mg	0	0
3	GA	pos	pre	70			X (dizziness)				0	0
4	GA	pos	post	80	X	X	X (headache)				1	1
5	GA	pos	post	80							0	0
6	GA	pos	post	75	X	X			X	Morphine 20 mg	0	0
8	GA	pos	post	90							0	0
12	GA	pos	post	125							0	0
13	GA	pos	post	115	X		X (dizziness)				0	0
18	GA	pos	post	55	X				X	Morphine 20 mg	6	6
19	GA	pos	post	125			X (dizziness)		X	Ketorolac tromethamine 60 mg	0	0
20	GA	neg	post	100							0	0
24	GA	pos	post	60			X (drowsiness)		X	Ketorolac tromethamine 60 mg	0	0
25	GA	pos	post	120					X	Ketorolac tromethamine 60 mg	0	0
26	GA	neg	post	105							0	0
27	GA	pos	post	115							0	0
29	GA	pos	post	80							0	4
30	GA	pos	post	90							0	0
33	GA	pos	pre	125	X				X	Ketorolac tromethamine 30 mg	0	0
34	GA	pos	post	120							0	0
38	GA	pos	pre	85							5	4
40	GA	neg	post	90				X		Ketorolac tromethamine 30 mg	0	0
45	GA	pos	post	135							5	0
46	GA	pos	pre	90							0	0
48	GA	pos	post	75							4	4
49	GA	pos	pre	95					X	Morphine 20 mg	5	5
50	GA	pos	post	70							4	4
52	GA	pos	post	95							0	0
54	GA	pos	post	50							0	0
55	GA	pos	post	75							3	0
56	GA	pos	post	95							0	0
57	GA	pos	post	140							0	0
62	GA	pos	pre	85							0	0
65	GA	pos	pre	115							0	0
67	GA	pos	pre	75				X		Morphine 20 mg	0	0
68	GA	pos	post	40							0	0

For neuropathic pain, at 3–6 or 12–24 months post-surgery, values above a score of 4 are highlighted in red.

DN4, Douleur Neuropathique 4; GA, general anesthesia with opioids; OfGA, opioid-free general anesthesia with locoregional blocks.

Together, these data indicated that OfGA was better tolerated, with a notably lower incidence of side effects, while it provided pain control at least as effectively as GA.

Next, we measured the levels of 44 inflammatory and immunomodulatory factors (full list in online supplemental table 1), in patients enrolled in the study for which also plasma samples were successfully collected (n=45; 24 in GA arm and 21 in OfGA arm) (figure 1). Many inflammatory cytokines, upregulated in the GA arm in a time-dependent manner, displayed a less pronounced increase in the OfGA group of patients, in line with the NLR results (figure 2). Interleukin 6 (IL-6), part of a well-known inflammatory pathway, the Janus kinase/Signal transducers and activators of transcription (JAK/STAT) pathway, frequently activated in BC³³ and whose upregulation correlates with metastatic spreading, in mice,³⁴ was strongly upregulated in the GA arm, both at 1 hour and 24 hours post-surgery, while it only slightly increased in the OfGA group (figure 3A, B). At 24 hours, IL-6 plasma levels were significantly lower in the OfGA arm than in the GA arm (figure 3A, B). Similar to IL-6, platelet-derived growth factor-AA (PDGF-AA) and PDGF-AB/BB, whose pathway hyperactivation is frequently associated with BC progression,³⁵ were significantly increased 24 hours from surgery in plasma of patients from the GA arm, while its levels remained totally unchanged in patients under OfGA (figure 3C–F). Other cytokines, either pro-inflammatory or, in some cases, anti-inflammatory, such as IL-1ra or IL-10, also displayed a time-dependent increase post-surgery, but did not differ in the two anesthetic approaches (online supplemental figure 1). Together, these results suggest that patients treated with OfGA displayed a reduced production of some critical pro-inflammatory and immunosuppressive cytokines.

Figure 3. Circulating levels of pro-inflammatory interleukin 6 (IL-6), platelet-derived growth factor-AA (PDGF-AA), and PDGF-AB/BB cytokines are lower in breast cancer patients treated with opioid-free general anesthesia (OfGA) compared to those treated with general anesthesia (GA). (A) Graph reports the plasma concentration of IL-6 in plasma from liquid biopsies (LB) collected from patients at different time points (LB1, basal pre-surgery; LB2, 1 hour after surgery; LB3, 24 hours after surgery) after administration of GA or OfGA, as indicated. (B) Graph reports IL-6 relative change, calculated over its basal levels in the pre-surgical sample (LB1), in GA and OfGA cohorts, as indicated. (C–F) Same as in A and B, but referred to PDGF-AA (C and D) or PDGF-AB/BB (E and F). In all box plots, boxes include the first and third quartiles, line indicates the median, and whiskers indicate the minimum and maximum data values. Outliers are reported at the extremities. Statistical significance was calculated with Wilcoxon test (A, C, E) or Mann-Whitney U test (B, D, F), as more appropriate. *p<0.050, **p<0.01, ***p<0.001 and ****p<0.0001. ns, not significant.

Discussion

Several studies suggested that the use of opioid drugs during general anesthesia may lead to transient postoperative immunosuppression, potentially exacerbating the systemic inflammatory response triggered by BC surgery. The increased expression of opioid receptors in various neoplastic tissues further suggests that opioids may directly influence the transcription of tumor-promoting factors, altogether highlighting the complex interplay between the pro-inflammatory, immunomodulating effects of opioids, and the biological pathways involved in tumor progression.^{25 26}

The potential of opioid-free anesthesia to reduce postoperative inflammation has been previously explored,⁶ with multiple clinical studies suggesting that regional anesthesia may improve outcomes in cancer patients by mitigating surgical stress, minimizing opioid exposure, while enhancing immune surveillance.^{27 28 36} Despite encouraging preclinical data and clinical observations, a recent large trial reported no significant difference in progression-free survival between BC patients receiving GA and those receiving OfGA.³⁰ However, this study had several limitations that may have masked potential benefits. These include partial use of opioids and volatile anesthetics in the OfGA group, the heterogeneity of the patient population—spanning various breast cancer stages and subtypes—and the absence of subgroup analyses, such as the influence of surgery duration on outcomes.³⁰ Additionally, the expected early peaks of recurrence around 18 months were not observed in either arm, further complicating the interpretation of the results. Together, these factors may have obscured the identification of patients’ subgroups who could derive greater benefit from an opioid-free perioperative intervention.^{31 32}

The debate on clinical impact is ongoing, but research on OfGA continues, emphasizing its advantages in minimizing opioid-related side effects and addressing opioid use concerns. We addressed the real potential of using opioid-free anesthesia coupled with regional blocks (OfGA) in managing pain and reducing post-surgical inflammation, comparing its effects to conventional general anesthesia with opioids (GA).

We used the NLR as an indicator of systemic inflammation and designed a randomized controlled trial to minimize confounding factors. We established strict enrollment criteria, excluding patients with comorbidities, pre-existing inflammatory states, or those on chronic analgesic therapy, particularly opioids. This approach allowed us to evaluate the postoperative effects of opioids accurately. Both groups were composed of BC patients undergoing the same surgery type (ie, external mammary quadrantectomy with sentinel lymph node biopsy), thus ensuring comparable surgical trauma. Surgery duration was very similar among the two groups, with a median of 92 min in the GA arm and 86 min in the OfGA one. This homogeneity in surgical and clinical variables supports the attribution of observed differences in outcomes to the anesthetic technique itself.

We collected blood samples at three time points, one before and two after surgery, and used NLR as a measure of post-surgical stress, reflecting systemic inflammatory and immune responses. An increased NLR has been proposed as a prognostic indicator, correlating with unfavorable outcomes in both oncological and non-oncological conditions.^{13–16 37 38} Our findings show that OfGA significantly attenuated the postoperative inflammatory response, as reflected in both NLR values and inflammatory cytokines like IL-6 and PDGF-AA (figures 2 and 3). Other cytokines, including anti-inflammatory mediators such as IL-10 and IL-1ra, also changed post-surgery but were not significantly influenced by anesthetic type (online supplemental figure 1).

An increased NLR is generally a negative prognostic indicator across various cancers, including BC, where it correlates with reduced survival.^{14 16} Inflammation fosters tumor progression by promoting angiogenesis, vascular permeability, and tumor cell proliferation, while suppressing immune defenses, including natural killer (NK) cells and T cell function. Although not conclusive, our findings support this paradigm: patients receiving GA showed higher postoperative levels of VEGF, a key pro-angiogenic factor, compared with those in the OfGA group (online supplemental figure 1). Other studies have also shown significant opioid-related reductions in NK cell numbers and function following surgery—an observation of particular concern given the critical role of NK cells in tumor surveillance.^{31 39–41} The potential beneficial effects of propofol have also been reported,⁴² but, since both groups in our study received propofol, any such effects would be expected to apply equally to both arms. A recent study investigating patients undergoing radical prostatectomy found no significant difference in postoperative NLR or recurrence-free survival between opioid-free and opioid-inclusive anesthesia.⁴³ However, the more extensive and longer nature of prostatectomy compared with quadrantectomy—associated with a substantially higher inflammatory burden—may have obscured any potential benefits of opioid avoidance in that context.

Another scarcely explored area is the impact of opioid-free or opioid-sparing approaches on postoperative pain outcomes, and whether they might carry unforeseen drawbacks. To address this, we evaluated analgesic needs in the immediate postoperative period (24 hours after surgery) and found that a greater number of patients in the GA group required rescue analgesia—including four who needed morphine. In contrast, only one OfGA patient required ketorolac, and none required morphine. For chronic pain, assessed using the DN4 questionnaire, five patients per group reported neuropathic pain at 3–6 or 12–24 months postoperatively, with no significant difference between groups. After surgery, these BC patients were all receiving adjuvant hormonal therapy, known to cause joint and nerve-related pain (table 2).

Despite the relatively small sample size and the simple surgical procedure (quadrantectomy), these results strongly suggest that a correctly performed locoregional anesthesiologic technique can provide better postoperative pain control and fewer side effects than GA with opioids.

While most prior research has focused on morphine, we studied fentanyl, one of the most commonly used opioids in surgical practice. The literature does not conclusively show that such low intraoperative doses of fentanyl (eg, ~150 μg) may significantly impact systemic inflammation. However, fentanyl has been shown to reduce NK cell activity in postoperative settings, echoing earlier in vitro and animal studies.⁴⁴ Studies exploring the effect of intraoperative use of fentanyl in humans are scarce, and certainly, isolating the effect of fentanyl from other variables (eg, regional blocks) is challenging. Notwithstanding this limitation, our trial is among the first to evaluate the removal of fentanyl in a randomized setting. While our findings suggest that fentanyl—like morphine and other opioids—may modulate postoperative inflammatory responses with potential implications for recovery, we cannot disentangle its specific effect from that of regional blocks.

In conclusion, this randomized controlled trial highlights the potential benefits of opioid-free anesthesia combined with regional blocks in reducing postoperative inflammatory and immunosuppressive responses, improving post-surgical pain symptoms, and minimizing side effects. Larger and highly controlled studies, including follow-up clinical data, are needed to determine whether OfGA is feasible on a broader scale and whether it may improve clinical outcomes after BC surgery.

Data availability statement

Data are available upon reasonable request.