Total intravenous or inhalational volatile anaesthesia and survival after colorectal cancer surgery: a Swedish national registry study
1 Department of Anaesthesia and Intensive Care, Vastmanland Hospital, Vasteras, Sweden
2 Centre for Clinical Research, Uppsala University, Vastmanland Hospital, Vasteras, Sweden
3 Department of Surgery, Vastmanland Hospital, Vasteras, Sweden
4 Epistat AB, Uppsala, Sweden
5 Department of Surgical Sciences, Anaesthesiology and Intensive Care, Uppsala University, Uppsala, Sweden
This article is accompanied by an editorial by Nunez-Rodrigue et al., Anaesthesia 2025; 80: 353–356.
Summary
Introduction
Retrospective studies suggest that inhalational volatile anaesthetic agents may contribute to an increased risk of metastasis and reduction in survival rates when used during cancer surgery. This relationship may vary between cancer types due to different tumour biology and differences in surgical procedures. This study aimed to investigate the relationship between the type of anaesthetic used for maintenance of anaesthesia (propofol or inhalational volatile anaesthetic agent) and survival in patients with stage 1–3 colorectal cancer who underwent resection surgery under general anaesthesia in Sweden between 2014 and 2019.
Methods
We identified patients from the Swedish Colorectal Cancer Registry. Their data, including cancer characteristics and adjuvant treatment, were then merged with information from the Swedish Peri-operative Registry. The primary outcome was overall survival, with disease-free survival as a secondary outcome.
Results
Of the 11,598 patients included, 8161 had colon cancer and 3437 had rectal cancer. General anaesthesia was maintained with propofol in 2346 (20%) patients, while 9252 (80%) received an inhalational volatile anaesthetic agent. In the unmatched cohort, patients who had general anaesthesia maintained with propofol for colon cancer surgery showed improved survival compared with those receiving an inhalational volatile anaesthetic agent (hazard ratio 0.83, 95%CI 0.72–0.95, p = 0.008). After 1:2 propensity score matching, we observed no significant difference in survival rates in this group (hazard ratio 0.89, 95%CI 0.76–1.04, p = 0.127). In the rectal cancer group, there was no difference in survival in either the unmatched cohort (hazard ratio 0.83, 95%CI 0.65–1.08, p = 0.166) or after propensity score matching (hazard ratio 0.95, 95%CI 0.71–1.25, p = 0.702). There was no significant difference in disease-free survival in either type of cancer.
Discussion
We found no association between the choice of agent for maintenance of anaesthesia and long-term survival outcomes in patients with colorectal cancer.
Introduction
Whether the choice of drug used for maintenance of general anaesthesia can affect long-term survival following cancer surgery remains a topic of debate [1-5]. Several underlying biological mechanisms have been proposed [6, 7]. Inhalational volatile anaesthetic agents are known to be pro-inflammatory [8]; potentially upregulate hypoxia-inducible factor-1α produced by tumour cells [9]; and impair DNA repair [10, 11]. Thus, inhalational volatile anaesthetic agents may theoretically exacerbate the inflammatory cascade initiated by major surgery, such as resection surgery for colorectal cancer, potentially promoting metastasis and recurrence.
In contrast, propofol has been suggested to have protective effects, which could possibly improve survival outcomes [12, 13]. Despite these findings, no difference in breast cancer survival between patients receiving sevoflurane or propofol for maintenance of anaesthesia was shown in a recent randomised trial [14]. However, it is plausible that a significantly shorter anaesthetic time for breast cancer surgery (compared with most other cancer surgeries) and differences in tumour biology between different forms of cancer could explain this result. In patients with colorectal cancer, two large registry studies comparing propofol and inhalational anaesthetic agents presented somewhat divergent results [15, 16]; one found no difference in either overall or disease-free survival between the study groups [15], whereas the other “found a weak association between recurrence and exposure to inhalational anaesthesia” [16]. One retrospective single-centre study looking at colon cancer indicated better overall survival with propofol compared with desflurane-based anaesthesia [5], while another small retrospective single-centre study investigating colorectal cancer found no difference in either overall or recurrence-free survival between patients receiving propofol or sevoflurane [17]. A recent systematic review indicated improved survival with propofol-based total intravenous anaesthesia but found no significant difference between total intravenous anaesthesia and inhalational volatile anaesthetic agents when studies from Europe and Asia were analysed separately [18].
Given the conflicting evidence, well-powered, population-based prospective studies with high validity, together with the results of ongoing randomised controlled trials, are needed to determine the relationship between anaesthetic agents and cancer survival. The aim of this study was to investigate the effect of the anaesthetic agent used for maintenance of anaesthesia (propofol or an inhalational volatile anaesthetic) on survival after surgery for stage 1–3 colorectal cancer, using pooled data from the Swedish Colorectal Cancer Registry and the Swedish Peri-operative Registry. We hypothesised that total intravenous anaesthesia with propofol would improve survival after colorectal cancer surgery.
Methods
After obtaining ethical approval, we used pooled data from the Swedish Colorectal Cancer Registry and the Swedish Peri-operative Registry. The Swedish Colorectal Cancer Registry contains a comprehensive data set for everyone diagnosed with colorectal cancer in Sweden, including: tumour stage; type of surgery; oncological treatment; cancer follow-up; and survival data with a coverage rate of 99% and high validity [19]. The Swedish Peri-operative Registry contains data on surgical procedures including which anaesthetic technique was used for induction and maintenance, but it does not discriminate between different inhalational volatile anaesthetic agents. The registry covers the entire peri-operative process from pre-operative assessment to discharge from the post-anaesthesia care unit with a coverage rate of 99% [20]. These registries rely on the unique social security number assigned to all Swedish citizens, they exhibit high data quality with built-in validation processes and have achieved the designation ‘quality registry’, which the government requires for funding. In the present study, we adapted the statistical plan from a previous study carried out to determine the effect of anaesthetic technique on outcomes after breast cancer surgery [21] (online Supporting Information Appendix S1).
All patients diagnosed with stage 1–3 colorectal cancer in Sweden who underwent surgery between 2014 and 2019, were identified from the Swedish Colorectal Cancer Registry, and their data were supplemented with pooled anaesthesia information from the Swedish Peri-operative Registry. In cases where a patient had multiple colorectal cancer sites, inclusion was based on the most advanced cancer site. For patients with a new cancer diagnosis during the follow-up period, we included the earliest resection operation. Patients with missing information were not studied.
The independent/causal variable was the anaesthetic drug used for maintenance, i.e. propofol or an inhalational volatile anaesthetic agent (desflurane, isoflurane or sevoflurane). The dependent/control variables included: age at surgery; BMI; and ASA physical status. We considered them to be true confounders because of their potential association with both the choice of anaesthetic agent and overall survival. Other factors such as tumour stage (1/2/3); year of diagnosis; neoadjuvant and/or adjuvant oncological therapy (chemotherapy, radiotherapy); and type of surgery (laparotomy, laparoscopy), were judged to be associated with prognosis, but not with the choice of anaesthesia, and were therefore included as effect modifiers in a propensity score matching model.
Colon and rectal cancers were analysed separately to account for differences in cancer biology, location and surgical techniques, which may affect the duration and timing of anaesthetic exposure. The primary outcome was overall survival, defined as the time from inclusion to death from any cause. The secondary outcome was disease-free survival, defined as the time from surgery to recorded recurrence or death.
Only patients with complete information were included in the analyses of overall survival. Overall survival was presented using the Kaplan–Meier method with the corresponding log-rank test. In addition, overall mortality, expressed as hazard ratios with 95%CIs, was estimated between the two groups using Cox regression models. We tested the proportional hazards assumption for all the Cox regressions using the tests based on weighted residuals. In the next step, we created two propensity score-matched cohorts (propofol/inhalational volatile anaesthetic). Propensity scores were developed, accounting for all patient and clinical characteristics summarised in Table 1. All individual propensity scores were calculated by logistic regression models, and then a 1:2 nearest neighbour propensity score matching with a calliper size of 0.1 was used.
| Propofol | Inhalational volatile | SMD | |
|---|---|---|---|
| n = 1658 | n = 6503 | ||
| Age; y | 74 (67–80 [23–95]) | 75 (78–81 [17–97]) | 0.031 |
| Sex; male | 831 (50%) | 3186 (49%) | 0.023 |
| Follow up time; months | 67 (41–103 [0–202]) | 76 (45–119 [0–205]) | |
| Calendar year | |||
| 2014–2015 | 176 (11%) | 1219 (19%) | 0.278 |
| 2016–2017 | 529 (32%) | 2310 (35%) | |
| 2018–2019 | 953 (57%) | 2974 (46%) | |
| ASA physical status | |||
| 1–2 | 1072 (65%) | 4044 (62%) | 0.051 |
| 3–4 | 586 (35%) | 2459 (38%) | |
| Laparoscopic surgery | 859 (52%) | 2666 (41%) | 0.218 |
| Location | |||
| Colon right | 1031 (62%) | 3950 (61%) | 0.030 |
| Colon left | 627 (38%) | 2553 (39%) | |
| Stage at diagnosis | |||
| 1 | 388 (23%) | 1273 (20%) | 0.111 |
| 2 | 680 (41%) | 2621 (40%) | |
| 3 | 590 (36%) | 2609 (40%) | |
| Surgery | |||
| Acute | 142 (9%) | 754 (12%) | 0.101 |
| Elective | 1516 (91%) | 5749 (88%) | |
| Volume of blood loss; ml | 50 (25–150 [0–5000]) | 50 (25–200 [0–12,000]) | 0.103 |
- SMD, standardised mean difference.
All tests were two-sided and statistical significance was considered with a p value < 0.05. The statistical analyses were performed using R version 3.6.1. (R basis for statistical calculation, Vienna, Austria).
Results
A total of 14,837 patients with stage 1–3 colorectal cancer, for whom information was available regarding the anaesthetic technique used, underwent tumour resection in Sweden between 2014 and 2019. Following exclusions, the study population comprised 11,598 patients: 8161 with colon cancer and 3437 with rectal cancer (Fig. 1).
In the colon cancer group, 1658 (20%) patients received propofol for maintenance of anaesthesia, while 6503 (80%) received an inhalational volatile anaesthetic agent. The proportions were similar in the rectal cancer group: 688 (20%) patients received propofol for maintenance and 2749 (80%) received an inhalational volatile anaesthetic agent. There were no significant differences in age, sex, ASA physical status or tumour stage between groups. Propofol was used more frequently than inhalational volatile anaesthetic agents in patients undergoing laparoscopic surgery for both colon and rectal cancer (Table 2). After propensity score matching, the cohort comprised 2346 patients who received propofol maintenance and 4686 patients who received an inhalational volatile anaesthetic agent (Tables 3 and 4).
| Propofol | Inhalational volatile | SMD | |
|---|---|---|---|
| n = 688 | n = 2749 | ||
| Age; y | 70 (62–76 [31–94]) | 70 (62–76 [23–93]) | 0.027 |
| Sex; male | 414 (60%) | 1701 (62%) | 0.035 |
| Follow up time; months | 82 (48–122 [0–109]) | 75 (45–109 [0–122]) | |
| Calendar year | |||
| 2014–2015 | 69 (10%) | 518 (19%) | 0.260 |
| 2016–2017 | 264 (38%) | 1022 (37%) | |
| 2018–2019 | 355 (52%) | 1209 (44%) | |
| ASA physical status | |||
| 1–2 | 521 (76%) | 2000 (73%) | 0.068 |
| 3–4 | 167 (24%) | 749 (27%) | |
| Laparoscopic surgery | 432 (63%) | 1486 (54%) | 0.178 |
| Stage at diagnosis | |||
| 1 | 241 (35%) | 876 (32%) | 0.098 |
| 2 | 207 (30%) | 786 (29%) | |
| 3 | 240 (35%) | 1087 (39%) | |
| Volume of blood loss; ml | 100 (500–300 [0–5800]) | 150 (50–400 [0–13,000]) | 0.188 |
- SMD, standardised mean difference.
| Propofol | Inhalational volatile | SMD | |
|---|---|---|---|
| n = 1658 | n = 3310 | ||
| Age; y | 74 (67–80 [17–96]) | 74 (67–80 [23–95]) | 0.005 |
| Sex; male | 831 (50%) | 1655 (50%) | 0.002 |
| Calendar year | |||
| 2014–2015 | 176 (11%) | 336 (10%) | 0.015 |
| 2016–2017 | 529 (32%) | 1066 (32%) | |
| 2018–2019 | 953 (58%) | 1908 (57%) | |
| ASA physical status | |||
| 1–2 | 1072 (65%) | 2138 (65%) | 0.001 |
| 3–4 | 586 (35%) | 1172 (35%) | |
| Laparoscopic surgery | 859 (52%) | 1698 (51%) | 0.010 |
| Location | |||
| Colon right | 1031 (62%) | 2046 (62%) | 0.008 |
| Colon left | 627 (38%) | 1264 (38%) | |
| Stage at diagnosis | |||
| 1 | 388 (23%) | 732 (22%) | 0.035 |
| 2 | 680 (41%) | 1355 (41%) | |
| 3 | 590 (36%) | 1223 (37%) | |
| Elective surgery | 1516 (91%) | 3039 (92%) | 0.014 |
- SMD, standardised mean difference.
| Propofol | Inhalational volatile | SMD | |
|---|---|---|---|
| n = 688 | n = 1376 | ||
| Age; y | 70 (62–76 [23–92]) | 71 (63–76 [31–94]) | 0.001 |
| Sex; male | 414 (60%) | 827 (60%) | 0.021 |
| Calendar year | |||
| 2014–2015 | 69 (10%) | 144 (10%) | 0.018 |
| 2016–2017 | 264 (38%) | 532 (39%) | |
| 2018–2019 | 355 (52%) | 700 (51%) | |
| ASA physical status | |||
| 1–2 | 521 (76%) | 1067 (77%) | 0.043 |
| 3–4 | 167 (24%) | 309 (23%) | |
| Laparoscopic surgery | 432 (63%) | 858 (62%) | 0.009 |
| Stage at diagnosis | |||
| 1 | 241 (35%) | 468 (34%) | 0.024 |
| 2 | 207 (30%) | 427 (31%) | |
| 3 | 240 (35%) | 481 (35%) | |
- SMD, standardised mean difference.
In the unmatched analysis, during a median (IQR [range]) follow-up of 76 (45–116 [0–205]) months, overall survival was longer in patients with colon cancer receiving propofol for maintenance of anaesthesia compared with those who received an inhalational volatile anaesthetic agent (hazard ratio 0.83, 95%CI 0.72–0.95, p = 0.008) (Table 5, online Supporting Information Figure S1). However, after 1:2 propensity score matching (1658 patients in the propofol group and 3310 patients in the inhalational volatile group) the difference in overall survival lost significance (hazard ratio 0.89, 95%CI 0.76–1.04, p = 0.127) (Table 5, Fig. 2).
| 1-year survival | 3-year survival | ||||
|---|---|---|---|---|---|
| Propofol | Inhalational volatile | Propofol | Inhalational volatile | Hazard ratio (95%CI) | |
| Overall survival; % | |||||
| Colon cancer, unmatched | 95.2 (94.1–96.2) | 93.6 (93.0–94.2) | 84.9 (82.8–87.0) | 81.8 (80.7–82.9) | 0.83 (0.72–0.95) |
| Rectal cancer, unmatched | 96.8 (95.4–98.1) | 95.6 (94.8–96.4) | 87.5 (84.5–90.5) | 87.8 (86.4–89.2) | 0.83 (0.65–1.08) |
| Colon cancer, propensity score matched | 0.89 (0.76–1.04) | ||||
| Rectal cancer, propensity score matched | 0.95 (0.71–1.25) | ||||
| Disease-free survival; % | |||||
| Colon cancer, unmatched | 94.9 (93.9–96.0) | 93.2 (92.6–93.8) | 84.4 (82.3–86.6) | 81.3 (80.2–82.4) | 0.83 (0.72–0.96) |
| Rectal cancer, unmatched | 96.5 (95.1–97.9) | 95.1 (94.2–95.9) | 86.2 (83.1–89.4) | 86.8 (85.4–88.3) | 0.84 (0.66–1.08) |
| Colon cancer, propensity score matched | 0.90 (0.77–1.05) | ||||
| Rectal cancer, propensity score matched | 0.95 (0.72–1.25) | ||||
For patients with rectal cancer, 688 received propofol for maintenance of anaesthesia and 1376 received an inhalational volatile anaesthetic agent. No difference in overall survival between the two groups could be shown, either in the unmatched cohort (hazard ratio 0.83, 95%CI 0.65–1.08, p = 0.166) (Table 5, online Supporting Information Figure S2), or in a 1:2 propensity score-matched cohort (hazard ratio 0.95, 95%CI 0.71–1.25, p = 0.702) (Table 5, Fig. 3). The 1- and 3-year survival estimates are also shown in Table 5.
Disease-free survival in the unmatched analysis favoured propofol in colon cancer patients (hazard ratio 0.83, 95%CI 0.72–0.96, p = 0.009) (Table 5, online Supporting Information Figure S3), but after 1:2 propensity score matching the difference lost significance (hazard ratio 0.90, 95%CI 0.77–1.05, p = 0.172) (Table 5, online Supporting Information Figure S4). In patients with rectal cancer, there was no significant difference in disease-free survival in either the unmatched cohorts (hazard ratio 0.84, 95%CI 0.66–1.08, p = 0.175) (Table 5, online Supporting Information Figure S5), or in the 1:2 propensity score-matched cohorts (hazard ratio 0.95, 95%CI 0.72–1.25, p = 0.709) (Table 5, online Supporting Information Figure S6).
Discussion
In this comprehensive national registry-based study of patients with stage 1–3 colorectal cancer, we found no association between the choice of anaesthetic agent for maintenance of general anaesthesia – either propofol or an inhalational volatile anaesthetic agent – and either overall or disease-free survival. This finding aligns with the results of Makito et al. [15] and Hasselager et al. [16], who reported a lack of significant association between the choice of anaesthetic agent and disease-free survival or overall survival after colorectal cancer surgery in registry-based studies. However, Hasselager et al. noted an increased risk of cancer recurrence with inhalational-based anaesthesia compared with intravenous anaesthetics [16]. In the study by Makito et al., uncertainties regarding the completeness of follow-up data and the relatively short follow-up periods may have affected the robustness of their findings [15].
Although our initial unadjusted analysis favoured propofol for improved survival in colon cancer patients, this association disappeared when adjusted with propensity score matching, suggesting that selection bias or unobserved confounding variables may have impacted the initial findings. These results are consistent with our initial retrospective single-centre registry-based study that included both breast and colorectal cancer [1].
Propofol has replaced barbiturates as the anaesthetic induction agent of choice in Sweden. Thus, most patients receiving an inhalational volatile anaesthetic agent for maintenance of anaesthesia have also received a bolus dose of propofol. If the hypothesis that propofol has anti-cancer properties is correct, it is possible that an induction dose of propofol is sufficient to reduce the proposed pro-carcinogenic properties of inhalational volatile anaesthetic agents. This could potentially mask a true difference in long-term survival between propofol and inhalational volatile anaesthetic agents. It has been shown by Schaefer et al. that an increasing dose of propofol was associated with reduced odds of 1-year mortality, although not for patients with a solid cancer, and this was replicated in a further analysis of 5-year mortality [22]. For patients with colorectal cancer, however, there were increased odds of 1-year mortality with higher doses of propofol [22]. Although it cannot be ruled out that an induction dose of propofol may to some extent offset for a negative effect of an inhalational volatile anaesthetic agent, we believe it is unlikely to have modified our results.
This study had several limitations. Despite the prospective data collection in the two registries, their high coverage and the completeness of follow-up data, the retrospective nature of the analysis within this study introduces inherent limitations together with the non-randomised anaesthesia selection. Although we adjusted the analysis for known confounders by 1:2 nearest neighbour propensity score matching, the matching approach was not possible for all subjects due to small samples in some of the subgroups. Additionally, there may still be unmeasured confounders. The registries do not collect data on lifestyle factors, socio-economic conditions, medications or other comorbid diseases than cancer.
In addition, the Swedish Peri-operative Registry lacks data on the specific inhalational volatile anaesthetic agents (desflurane, isoflurane or sevoflurane) that were used for individual patients. Notably, sevoflurane dominates the market in Sweden, but since sales results are not public, we do not know to what extent. To our knowledge, there are only two studies comparing long-term survival between different inhalational volatile anaesthetic agents. Elias et al. showed in a small retrospective, single-centre study that the recurrence rate after cyto-reductive surgery for stage 3 ovarian cancer was significantly lower with desflurane compared with sevoflurane [23]. In another slightly larger, retrospective, single-centre study by Liu et al., isoflurane and sevoflurane were used for colorectal cancer surgery [24]. Extrapolated from a Kaplan–Meier graph in the publication, it appears that overall survival was almost 20 percentage points higher with sevoflurane. Although results from small single-centre studies must be interpreted with caution, it may be that desflurane, isoflurane and sevoflurane differ in their immunological properties. If so, our results become more difficult to interpret.
The existing literature highlights the urgent need for well-designed randomised trials in patients with colorectal cancer to determine the potential effects of anaesthetic technique on cancer outcomes. Pre-clinical data support the hypothesis that propofol has anti-tumorigenic properties and that inhalational volatile anaesthetic agents are pro-tumorigenic, and a majority of retrospective clinical studies indicate a survival benefit in favour of propofol; these findings should ideally be tested in randomised clinical trials [7]. However, given the nature of the outcome measure being analysed, few trials are to be expected and large retrospective observational studies based on prospectively collected data may be the best alternative.
In conclusion, this study using merged data from two national quality registries, revealed no discernible association between the choice of anaesthetic agent for maintenance of anaesthesia at surgery and long-term survival outcomes in patients with colorectal cancer. Pending results from randomised studies, we find no basis for advocating a modification of local protocols regarding the risk of cancer recurrence based on the choice of anaesthetic agent.
Acknowledgements
The data used in this study were taken from the Swedish colorectal cancer register and the Swedish peri-operative register. The registries take no responsibility for the methods, analyses or results presented here. The authors express their gratitude to the steering groups of the Swedish Colorectal Cancer Registry and the Swedish Peri-operative Registry and to their administrators for great support, especially Beata Pajak and Camilla Hartmann-Norman at the Uppsala Clinical Research Center (Uppsala, Sweden). ME participates in the ESAIC Onco Anesthesiology Research Group, EuroPeriscope and is grateful to this organisation. Anonymised individual patient data collected for this study can be made available on request via the corresponding author, considering possible legal restrictions (Swedish and EU; GDPR) and after presenting a sound proposal and ethics approval. A signed data access agreement is mandatory, including a description of the conditions for data release and the requirements for data transfer, storage, archiving, publication and co-authorship. A study protocol, statistical analysis plan and informed consent form may be asked for. The time frame for data sharing of colorectal cancer patients is defined from six months after publication of the current manuscript up to three years thereafter. Statistical code is not available. AB has received honoraria for arranging a course in cooperation with AstraZeneca. ME has received consulting fees from Nimbelle AB, honoraria for lectures from Malardalen University, is on the board of the ENCORE trial and is the initiator and the head PI for the CAN study. No external funding or other competing interests declared.
