In many parts of the world, cancer, together with cardiovascular disease, is now the number one cause of death. Furthermore, its incidence is increasing, with 14.1 million new cases of cancer diagnosed worldwide in 2012. Lung, breast, bowel, and prostate cancer are currently the most prevalent types.1 Importantly, mortality in many cancer patients is commonly the result of recurrence or metastasis rather than an effect of the primary tumour itself.

It is almost a decade since the hypothesis was put forward that perioperative factors, including anesthetic and analgesic management, may influence cancer recurrence or metastasis.2,3 Earlier still, breast cancer patients who had received halothane anesthesia showed increased survival rates when compared with those who had received ether anesthesia.4 A further direct effect of anesthetic agents on cancer cell biology was proposed in 1981 when the effect of four commonly used anesthetic drugs was examined on the postoperative growth of mouse tumours. This work showed accelerated growth of lung metastases and the development of secondary tumours in organs not otherwise commonly associated with metastases.5

The challenge of conducting large multicentre randomized-controlled trials that address this question cannot be underestimated. The main limitation of the majority of work published to date is its retrospective design. There are frequent uncontrolled and unrecognized biases and differing criteria regarding such data as the stage of cancer at the time of surgery, the underlying tumour biology, the surgical skill of the clinicians, and the effects of perioperative adjuvant therapies such as chemotherapy or radiotherapy.6,7

This present narrative review summarizes the current state of evidence for many regularly used drugs and techniques for anesthesia and analgesia in the perioperative period and addresses their actual or potential effects on cancer cell biology and the risk of cancer metastasis.

The literature in this article was retrieved from a search of the PubMed database up to May 31, 2015. Search terms included “anesthesia and metastases”, “anesthetic drugs and cancer”, “volatile anesthetic agents and cancer”, and “anesthetic technique and cancer”. Results were restricted to the English language. Relevant articles were also obtained from the articles identified in the literature review, and all primary sources were retrieved. Two authors then reviewed these articles for suitability.

Current laboratory research points to the existence of signals suggesting that perioperative interventions may impact recurrence or metastasis through effects on cancer cell signalling, the immune response, and modulation of the neuroendocrine stress response. Meanwhile, we await substantiated evidence from prospective randomized-controlled trials (RCTs) (Table). Understanding the molecular components and processes involved in metastatic and local spread may lead to a more targeted approach to prevent cancer recurrence as a consequence of perioperative management.8 Several animal model studies have shown that tumour growth and metastasis are enhanced by a combination of processes that occur in the perioperative period, including inadvertent dispersal of tumour cells during surgical manipulation, the suppression of cell-mediated immunity by the stress response, and the increased expression of proangiogenic factors by tumour cells that are an essential component of metastatic spread.7

Table Ongoing research investigating the effects of anesthetic agents on immune cell function and metastasis

The recent “Consensus statement from the BJA Workshop on Cancer and Anaesthesia” acknowledges insufficient evidence to support a change in clinical practice and calls for further research, preferably in the form of RCTs, to investigate the effects of volatile agents, opioids, nonsteroidal anti-inflammatory drugs (NSAIDS), and regional anesthetic techniques on cancer recurrence and metastasis.9 Indeed, although the evidence is inconclusive, the possibility that anesthesia may contribute to the recurrence of cancer cannot be ignored and requires definitive prospective RCTs to test whether a link of cause and effect exists between the most promising perioperative interventions and the risk of recurrence.

Are volatile anesthetic agents safe to use in cancer patients?

This question arises because an increasing number of studies have investigated the direct effects of volatile anesthetic agents on cancer cell biology, and results have shown that they induced mitogenesis, angiogenesis, and metastasis in tumours. One such study investigated the effects of isoflurane on the expression of tumourigenic markers, including insulin-like growth factor (IGF-1) and proliferative capacity in ovarian cancer cells. The study results showed that isoflurane significantly increased expression of IGF-1 and IGF-1 receptors, cell cycle progression, and cell proliferation in ovarian cancer cells 24 hr after exposure. It also showed increased expression of the angiogenic markers vascular endothelial growth factor (VEGF) and angiopoietin-1 after isoflurane exposure together with an enhanced angiogenesis. Enhanced cell migration of cancer cells after isoflurane exposure was associated with increased production of metalloproteinases 2 and 9, enzymes that play a key role in the degradation of extracellular matrix and thus facilitate local spread of tumour cells. This effect was terminated when IGF-1 signalling was blocked either by a neutralizing antibody or by small interfering ribonucleic acids.10 On this theme, results of a small RCT among 40 patients presenting for colon cancer surgery showed that serum levels of proangiogenic factors VEGF-C and transforming growth factor beta-1 increased significantly in the patients who received volatile-based anesthesia vs propofol-epidural anesthesia.11

Stem cells possess the ability to self-perpetuate and give rise to mature cells of a particular tissue through differentiation.12 Cancer stem cells are relatively rare cells with indefinite proliferative potential that drive the formation and growth of tumours.13 Studies investigating the effect of volatile agents on a range of cancer stem cells have shown varying results. One such study evaluating the response of glioma stem cell exposure to varying durations and concentrations of sevoflurane compared with controls showed increased proliferation of cancer cells and a capacity for self-renewal after sevoflurane. This was associated with increased expression of hypoxia-inducible factor (HIF) protein.14 Similarly, HIF was the mechanism underpinning the enhanced migration effect of isoflurane on renal cancer cells in vitro.15

In a recent study by Jaura et al., breast cancer cells were exposed to serum from females (n = 20) who underwent surgery for biopsy-proven primary breast cancer. The women were enrolled in a multicentre prospective clinical trial ( number, NCT00418457) and randomized to receive either propofol-paravertebral anesthesia or a volatile agent and opioid-based anesthetic technique for breast cancer excision. Study results showed less cancer cell apoptosis in the group receiving propofol-paravertebral anesthesia.16

Kvolik et al. investigated the cytotoxic and antiproliferative effects of sevoflurane in clinically relevant concentrations and duration of exposure on different human cancer cell lines, colon adenocarcinoma cells, and laryngeal carcinoma cells, in vitro. 17 They found that the apoptotic ratio had increased significantly 24 hr after anesthesia and was associated with increased expression of p53 and caspase-3 genes in colon cancer cells. Nevertheless, they also noted a decreased expression in laryngeal cancer cells, suggesting any potentially beneficial effect of this volatile agent to increase cancer cell apoptosis may be dependent on the tumour cell line.

On balance, while there remains conflicting evidence regarding the potential deleterious effects of volatile agents from the available evidence of in vitro studies to date, there is insufficient evidence to justify avoidance of these agents in cancer patients.

The immunosuppressive effect of nitrous oxide, mediated through selective inhibition of methionine synthase and therefore DNA, purine, and thymidylate synthesis, causes depressed macrophage and natural killer (NK) cell function.18,19 The recent ENIGMA-II trial found that use of nitrous oxide did not increase major postoperative complications at 30 days, including myocardial infarction, and there was no evidence of an aggravating effect relating to cancer recurrence or mortality.20 Prior to the ENIGMA-II trial, a study with a specific focus on nitrous oxide and cancer assessed the recurrence rate of colon cancer in a prospective randomized trial of 204 patients assigned to either 65% nitrous oxide or nitrogen during surgery. This study found a similar rate of recurrence in the two groups, suggesting that nitrous oxide has no impact on cancer recurrence.21

On the subject of specific anesthetic gases, the noble gas xenon has been tested clinically and found to have many desirable properties, but its progression into routine clinical practice has been limited in part by its prohibitive cost. Recent in vitro studies conducted on its effects on breast cancer cell biology found that acute exposure of breast adenocarcinoma cells to xenon inhibited migration without affecting viability and was associated with decreased secretion of pro-angiogenic cytokines RANTES/CCL5.22 Notably in this study, effects of xenon were visible within one hour of exposure of cancer cells to xenon. Although the mechanism by which xenon affects cancer biology is yet to be fully elucidated, the authors of this study suggest that the action of xenon at the glycine binding site of the N-methyl-D-aspartate receptor causes a decrease in breast cancer cell migration. Xenon is also thought to bind to the Kir6.2 subunit of the adenosine triphosphate-dependent potassium channel, causing cytoskeletal changes that are involved in tumourigenesis and cell migration.23

Local anesthetics and neuraxial anesthesia

Some retrospective clinical studies have found an association between the use of regional anesthesia and reduced cancer metastasis. Regardless of the local anesthetic agent employed, regional anesthetic techniques are associated with decreased activation of the stress response.24 This allows for a reduction in opioid usage as well as decreased release of endogenous opioids25 and, consequently, less immunosuppression. It has also been proposed that local anesthetic agents per se reduce the incidence of cancer recurrence through anti-inflammatory action as well as direct effects on the proliferation and migration of cancer cells.26 Both lidocaine and ropivacaine have been shown to have antiproliferative effects on cancer cells in vitro. 27 While these drugs are known to act via sodium (Na+) channel inhibition, their potentially beneficial effect in cancer cells is mediated by the alternative mechanism via inhibition of the proto-oncogene Src. The functions of Src include regulation of cell-to-cell adhesion and involvement in fibroblast cell division, and its overexpression in solid tumours is associated with tumour cell progression, invasion, and metastasis. The inhibition of Src thereby plays a role in preventing the release of downstream mediators of cancer cell migration and metastasis. Lidocaine, in clinically relevant plasma concentrations, has been shown to suppress proliferation of the cancer cell via a direct inhibitory effect on the epidermal growth factor receptor, which is a tyrosine kinase receptor essential for the proliferation and differentiation of epithelial cells and tumours of epithelial cell origin.28 Amide local anesthetics have also been shown to be potent inhibitors of voltage-activated Na+ channels and in this way inhibit metastatic colon cancer cell invasion.29

At clinically relevant concentrations, lidocaine also has an action on the methylation in tumour cells. The demethylation of DNA in breast cancer cells is associated with decreased tumour progression through activation of tumour suppressor genes.30 Lirk et al. found that lidocaine and, to a lesser extent, ropivacaine and bupivacaine exert demethylating effects on breast cancer cells in vitro.31 Furthermore, lidocaine, in combination with the chemotherapeutic agent 5-aza-2’-deoxycytidine, exhibits additive demethylating effects.31

Therefore, existing evidence from laboratory cell culture models clearly suggests that amide local anesthetics have a direct therapeutic effect of inhibiting cancer cell metastasis. To date, there is a lack of in vivo animal model data evaluating this hypothesis. It also seems warranted to conduct a prospective RCT that tests the effects of administering intravenous lidocaine during primary cancer surgery on long-term cancer outcome.

Studies concerning the use of neuraxial or regional anesthesia in cancer surgery have shown conflicting results with respect to the association of reduced cancer recurrence with cancer spread.32,33 In addition to the demethylating effects of local anesthetics, another principal mechanism by which neuraxial anesthesia might reduce cancer spread is via either an opioid sparing effect or attenuation of the surgical stress response and its immunosuppressive consequences. Compared with general anesthesia alone, the addition of spinal blockade to general anesthesia was shown to attenuate metastasis significantly in rats inoculated with a strain of breast adenocarcinoma. Furthermore, this effect was inversely related to NK cell expression, i.e., metastasis was reduced when NK cell expression increased.34 A recent large (n = 3,284) retrospective analysis examined the long-term outcomes (e.g., systemic cancer progression, recurrence, prostate cancer mortality, and all-cause mortality) of patients who underwent radical retropubic prostatectomy for adenocarcinoma. Their results showed general anesthesia alone was associated with an increased risk for systemic progression (hazard ratio [HR], 2.81; 95% confidence interval [CI], 1.31 to 6.05; P = 0.008) and higher all-cause mortality (HR, 1.32; 95% CI, 1.00 to 1.74; P = 0.047) compared with the combination of general and neuraxial anesthesia. There was also a trend towards an increased risk of death from prostate cancer in those patients who received only general anesthesia (adjusted HR, 2.2; 95% CI, 0.88 to 5.60; P = 0.091).35

Nevertheless, although there are multiple retrospective analyses evaluating regional anesthesia techniques in many kinds of cancer, there are conflicting results.36 A recent translational study conducted on a subset of patients enrolled in an ongoing multicentre RCT in breast cancer patients found that the serum of women (n = 10) with breast cancer randomized to receive propofol-paravertebral anesthesia for excisional surgery preserved NK immune cell cytotoxicity and breast cancer cell apoptosis to a greater extent than the serum of women randomized to receive standard volatile-based general anesthesia with opioid analgesia.37 Practitioners may choose to promote regional anesthesia with propofol-based general anesthesia to their cancer patients for many reasons other than the possibility of an effect on cancer recurrence. The continuing prospective RCT ( number, NCT00418457) may prove to address whether this technique has a cause and effect link to cancer recurrence.

Myles et al. conducted a follow-up subgroup analysis of the MASTER study from 200038 in an effort to identify whether performing an epidural block at the time of major abdominal surgery for cancer was associated with reduced cancer recurrence and improved survival. Of the 915 patients enrolled in the MASTER trial, 446 patients were included in the subgroup analysis, 216 did not receive an epidural, and 230 received an epidural at the time of their surgery. The results showed that the median time to recurrence of cancer or death was 2.8 yr (95% CI, 0.7 to 8.7) in the control group and 2.6 yr (95% CI, 0.7 to 8.7) in the epidural group (P = 0.61). Recurrence-free survival was similar in both the epidural and the control group (HR, 0.95; 95% CI, 0.76 to 1.17; P = 0.61).39

Nonsteroidal anti-inflammatory agents

It is suggested that the effect of NSAIDs on disease progression may be tumour-specific and mediated through a direct effect on cyclooxygenase-2 (COX-2) receptors expressed by cancer cells as well as by indirect antagonism of the production of prostaglandins by NK cells. The expression of COX-2 in many types of cancer cells is associated with an increased resistance to apoptosis and the production of prostaglandins involved in tumour growth.40 Both chronic and acute use of NSAIDs as analgesics in the cancer patient have been associated with tumour regression, presumably through the inhibition of COX-2 and prostaglandin synthesis, although alternative COX-independent pathways may also play a role in COX-2 selective NSAIDs.41-43 The effects of celecoxib, a selective COX-2 inhibitor, and SC-560, a selective COX-1 inhibitor, on cell survival, cell cycle distribution, and apoptosis in colon cancer cell lines were evaluated in vivo and in vitro. Results showed that both drugs induced a G0/G1 phase block and reduced cell survival. These effects were independent of COX-2 expression by the cancer cell.41 In another study comparing the effects of celecoxib and rofecoxib on human prostate cancer cell lines—which, notably, did not express COX-2—celecoxib was found to suppress the growth of these cell lines. Treatment of these cell lines with celecoxib led to a marked decrease in the expression of the protein-coding gene, cyclin D1, which is essential for cell cycle progression to synthesis or during cell DNA replication.42

Regarding acetylsalicylic acid, a meta-analysis of five large RCTs that randomized patients to either acetylsalicylic acid 75 mg daily vs control for the prevention of vascular events found that allocation to acetylsalicylic acid reduced risk of cancer with distant metastasis (all cancers: HR, 0.64; 95% CI, 0.48 to 0.84; P = 0.001; adenocarcinoma: HR, 0.54; 95% CI, 0.38 to 0.77; P < 0.001; other solid cancers: HR, 0.82; 95% CI, 0.53 to 1.28; P = 0.39), mainly due to a reduction in the proportion of adenocarcinomas that had metastatic vs local disease (odds ratio, 0.52; 95% CI, 0.35 to 0.75; P < 0.001).43

In a retrospective analysis of 720 patients conducted by Forget et al., the intraoperative use of ketorolac or diclofenac in breast cancer patients was associated with prolonged disease-free survival and overall survival.44 In the same group of patients, a raised neutrophil:lymphocyte ratio in the preoperative period was associated with reduced survival. These results differ from a similar retrospective study among 1,111 patients who underwent radical prostatectomy for localized prostate cancer. Findings indicated that the perioperative use of ketorolac failed to show any survival benefit.45

Currently available evidence suggests merit in using perioperative NSAIDs to lessen the risk of cancer recurrence; however, individual patient decisions whether or not to use them should be based on an individual risk-benefit assessment of their likely value in a particular case. Proof of any effect on cancer recurrence must await results of prospective RCTs.


Opioids remain a cornerstone in the management of both acute postoperative pain and cancer-related pain in palliative care. Nevertheless, there is a strong suggestion from both retrospective clinical trials and experimental studies that opioids may promote cancer progression and reduce long-term survival.

Opioids have been shown to have numerous direct effects on cancer cells, which culminate in facilitating the spread of malignant cells. Proposed mechanisms include the direct promotion of cancer cell growth and inhibition of cellular immunity. Opioids are capable of stimulating angiogenesis partly by activating COX-2. This increases the production of prostaglandin E2, which promotes angiogenesis and directly affects tumour growth.46 Morphine contributes to the proliferation and survival of cancer cells by activating mitogen-activated protein kinase and Akt signalling pathways in tumour cells.47

Several studies have focused on the actions of the μ-opioid receptor (MOR) in the regulation of tumour growth and metastasis.48,49 The overexpression of MOR has been shown in both lung50,51 and prostate cancer. 52 The theory regarding MOR involvement in metastasis is further supported by the diminished progression of lung carcinoma in MOR knockout mice treated with the opioid receptor antagonist naltrexone.53 Morphine does not influence the initiation of tumour development; rather, it affects the progression of established breast tumour, as recently shown in a transgenic mouse study that found significantly decreased survival in a mouse model of breast adenocarcinoma.54 In the same study, it was found that morphine promoted lymphangiogenesis, mast cell activation and degranulation, and increased levels of inflammatory cytokines, tryptase, and substance P. In addition, the morphine-treated mice had increased tumour burden and decreased duration of survival.54

In an immunohistochemical study on the effect of MOR on metastasis, patients who had excision of non-small cell lung carcinoma were evaluated for their MOR status. There was a direct correlation between MOR expression and the extent of metastasis in these patients.48

A further study carried out on the aforementioned breast cancer specimens from patients previously enrolled in the ongoing RCT ( number, NCT00418457) has investigated the effect of anesthetic technique on immune cell expression in breast cancer tissue. It was found that women with breast cancer randomized to receive propofol-paravertebral anesthesia (n = 12) for surgical excision had a higher expression of NK cell infiltration than women with standard vapour general anesthesia with opioid analgesia (n = 16). This finding suggests that anesthetic technique may have an early effect on immune function in breast cancer tissue.55

Supplemental oxygen effects

Excess oxygen administration has been associated with increased morbidity and mortality across a range of medical conditions.56,57 Several studies in postoperative patients treated regularly with high fractional inspired oxygen concentrations have shown varying results, particularly in relation to the development of surgical site infections in the postoperative period.58,59 In a retrospective follow-up of a previous RCT evaluating the effect of different concentrations of oxygen administered during primary cancer surgery, the incidence of new or recurrent cancers was similar after 80% or 30% oxygen [140 of 678 patients (21%) in the 80% group vs 150 of 699 patients (21%) in the 30% group]. Nevertheless, cancer-free survival was significantly shorter in the 80% oxygen group (HR, 1.18; 95% CI, 1.01 to 1.42; P = 0.04), and the median time between surgery and new cancer diagnosis was also significantly shorter (80% group, 335 days vs 30% group, 434 days; P = 0.047).60 As well as promoting the production of erythropoietin and new vessel formation,61 it is possible that supplemental oxygen at the time of surgery has a proangiogenic effect on micrometastases, which, in the absence of hyperoxia, may have otherwise remained quiescent. Oxygen therapy for prolonged periods is also associated with the formation of reactive oxygen species and DNA damage.62


Steroids administered in the perioperative period have immune-modulating effects that include suppression of NK cells and enhanced resistance of the tumour cell to apoptosis.63 Dexamethasone is an antiemetic frequently used in the perioperative period, and it may potentially be pro-analgesic.64 A recent study reporting the five-year follow-up of patients who received dexamethasone or placebo before elective colectomy for colon cancer as part of a previous RCT65 found no difference in overall or disease-free survival between the two groups. However, patients who had received a dose of dexamethasone (8 mg) had a significantly higher rate of distant recurrence compared with those who received placebo (six patients in the dexamethasone group vs one patient in the placebo group; P = 0.04) and a non-significant trend towards higher cancer-specific mortality.66

Blood transfusion

Over at least the past decade, the practice of blood transfusion has been influenced by the principle of avoidance until an anemia trigger of < 7 g·dL−1 has been reached. Nevertheless, results of a recent randomized trial (n = 198) of postoperative cancer patients, where patients were randomized to receive either a liberal transfusion regimen (transfused if hemoglobin [Hb] < 9 g·dL−1) vs a restrictive regimen (transfused only if Hb < 7 g·dL−1), showed that the composite (95% CI) of morbidity and mortality at 30 days after surgery for abdominal cancer was 19.6% vs 35.6%, respectively (95% CI, 12.9 to 28.6 in the liberal regimen vs 95% CI, 27 to 45.4 in the restrictive regimen; P = 0.012). This suggests that a more liberal transfusion regimen may be beneficial, at least in terms of early postoperative morbidity.67 This does not affect the long-term association between transfusion and increased cancer recurrence.68


In conclusion, a major limitation of available evidence is the fact that the in vitro conditions of most studies do not faithfully replicate cancer cell biology conditions in vivo. Therefore, there is a high degree of speculative extrapolation of results to the clinical setting. Although prospective RCTs (Table) are ongoing and we await their results, there is also a growing recognition that choice of anesthetic technique and agent at the time of cancer surgery may have significant long-term implications in tumour growth and spread.

With increasing numbers of patients presenting for cancer surgery and increasing signals suggesting possible long-term effects of anesthetic technique on cancer growth, there is a growing need to fund prospective multicentre RCTs that can address these questions with more certainty. Meanwhile, anesthesiologists and their patients should seek optimal anesthesia and analgesia based on individual risk-benefit analysis and best available evidence on outcomes other than cancer recurrence. Such decisions should be made on the basis of established evidence around optimal analgesia, avoidance of unwanted side effects, or compatibility with a patient’s comorbidities, rather than on any putative benefit in cancer patients.