Introduction

High-grade gliomas are notorious for a high recurrence rate even after gross curative surgical resection, possibly because intense stress and inflammatory response may occur during surgery, inducing immunosuppression that promotes locoregional cancer recurrence1. In addition, anesthesia-related factors such as the use of inhalation anesthesia and administration of opioids also induce detrimental immunosuppression2. By contrast, regional anesthesia effectively relives surgical stress, reduces inhalational anesthetic and opioid consumption, and has been reported to be beneficial to patients undergoing cancer surgery1,3. Therefore, there is increasing interest in investigations of the potential oncological benefits of regional anesthesia for cancer surgery4. To date, previous reports have indicated conflicting results regarding the influence of regional anesthesia, namely the scalp block, on postoperative glioma recurrence5,6,7. For instance, we previous reported that scalp block is associated with a prolonged postoperative glioma recurrence5. By contrast, another larger cohort revealed negative results6. This may be because a tumor’s genetic profile plays a decisive role in the oncological outcomes of patients with glioma8, but this has not been analyzed and compared in previous reports of scalp block for glioma resection.

The revised 2016 World Health Organization classification of central nervous system tumors reclassified gliomas on the basis of molecular marker diagnostics combined with classical histological diagnosis8. Possibly the foremost molecular markers are mutations in isocitrate dehydrogenase (IDH)9. Mutations in IDH isozyme 1 (IDH1) not only play a crucial role in early tumorigenesis of astrocytomas and oligodendrogliomas but are also the decisive genetic signposts of secondary glioblastoma10. Most IDH1 mutations reported are located at the first or second base of codon 132 and the most frequent is R132H (IDH1R132H)10. Patients with glioma and IDH1R132H glioblastoma have twice the median overall survival (OS) and more favorable progression-free survival (PFS) than those with the IDH1 wild type11,12. However, occurrence of IDH1R132H has not yet been analyzed in the literature regarding scalp block and oncological outcomes. Therefore, in this study, we investigated the influence of scalp block on oncological outcomes in patients undergoing resection surgery for high-grade glioma; a complete IDH1R132H genetic profile analysis was performed. To our best knowledge, this presents the first literature of scalp block and the oncological outcomes of high grade glioma with the analysis of IDH1R132H mutation.

Methods

After obtaining the local ethic committee approval, a prospectively maintained database including patients of high-grade glioma undergoing craniotomy for glioma curative resection with complete IDHR132 mutation information was used for analysis. Exclusion criteria were (1) age younger than 20 years, because molecular characteristics and prognosis may differ between pediatric and adult patients with high grade glioma13, (2) undergoing an awake craniotomy, and (3) receiving chemotherapy or radiotherapy before surgery. Data collected included demographic characteristics, comorbidities, tumor location, tumor size, perioperative red blood transfusion, amount of opioid consumed, presence of the IDH1R132H mutation, postoperative adjuvant therapies, and anesthetic techniques such as the use of intravenous or inhalation anesthesia and scalp block.

Surgery and anesthesia

Each patient received general anesthesia that was maintained either intravenously or by inhalation at the attending anesthesiologist’s discretion. Blood transfusion was performed when the patient’s hemoglobin level was less than 9 g/dL. In patients receiving inhalational anesthetics, sevoflurane was used and was kept at 0.5 times the minimum alveolar concentration to minimize interference with intraoperative electrophysiological monitoring14. Perioperative steroids were not routinely given because brain lymphoma sometimes resembles glioblastoma on preoperative magnetic resonance images15, and steroids administered perioperatively would interfere with pathological examination of the tumor. Application of scalp block was at the anesthesiologist’s discretion and was performed immediately after induction of general anesthesia and before head pinning with a Mayfield head holder; the scalp block regimen was 10 mL of 0.5% levobupivacaine and 1:200,000 epinephrine mixture for each side of the scalp. The effect of the scalp block was achieved by means of local infiltration around targeted nerves, which were the supraorbital, supratrochlear, zygomaticotemporal, auriculotemporal, greater occipital, lesser occipital, and least occipital nerves.

Following surgery, all patients were sent to the same neurosurgical intensive care unit for postoperative care. Postoperative analgesia was achieved with mild to moderate opioids such as tramadol or nalbuphine; nonsteroidal anti-inflammatory drugs were used once the risk of bleeding was considered minimal by the surgeons.

Statistical analysis

PFS was defined as the interval between the date of surgery and the date of first evidence of tumor recurrence, which was based on postoperative magnetic resonance imaging once every 3 months, as is recommended by guidelines followed by many hospitals worldwide16. OS was defined as the interval between the date of surgery and date of death or loss of follow-up. To identify the targeted PFS improvement for at least one follow-up interval (3 months), 90 patients would be needed for statistical analysis with a power of 0.8 and α = 0.05.

Fisher’s exact test or the chi-square test was employed to analyze dichotomous data, Student’s t-test was used for normally distributed continuous data, and the Mann–Whitney U test was used for nonparametric ordinal data. PFS and OS are presented as the median (95% confidence interval [CI]) and were calculated using Kaplan–Meier survival analysis with the log-rank test to compare survival curves between groups. Univariate and multivariate Cox proportional hazards regression analyses were used to determine the effect of several risk factors on PFS and OS. A p < 0.05 was considered significant.

In our institute, the complete information regarding the IDHR132 mutation was routinely reported in postoperative pathological records since 2014. Because the incidence of high grade glioma was approximately only one per 100,00017, it is difficult for patient enrollment. Therefore, this study included high grade glioma patients enrolled in our previous study, which include our institutional glioma patients during 2010 to 201718, to achieve a sufficient study power. Accordingly, a parallel multivariate Cox regression analysis without overlapping was also performed to investigate the influence of scalp blocks on the oncological outcomes of patients with high grade glioma. Statistical analysis was conducted using IBM SPSS Statistics for Windows, version 21.0 (IBM Corp., Armonk, NY, USA) and MedCalc Statistical Software version 18.2.1 (MedCalc Software bvba, Ostend, Belgium).

Ethics statement

This study was approved by the Research Ethics Committee of National Taiwan University Hospital (approval No.201911066RIND, approval date: December 27, 2019), and the informed consent requirement was waived by the committee. The patients’ identifying information was omitted during analysis. This study adheres to the applicable EQUATOR guidelines.

Results

Baseline characteristics

In total, 460 patients underwent craniotomy for tumor resection between January 1, 2014, and December 31, 2019; 112 were finally included in the analysis (Fig. 1). Seventeen patients were revealed to carry the IDH1 R132H mutation (15.2%). The median PFS and OS of all patients were 12.40 (8.63–15.17) and 36.83 (22.50–61.13) months, respectively. The 1-year PFS rate of all patients was 50.9%, whereas the median follow-up interval was 15.88 (14.33–18.79) months.

Figure 1
figure 1

Patient inclusion.

Full size image

Baseline characteristics between the patients receiving and not receiving scalp block were comparable (Table 1). In our institute, scalp block is often combined with total intravenous general anesthesia (p < 0.0001). Lower intraoperative fentanyl consumption was observed in the scalp block group (median fentanyl consumption: 250 [150–300] vs. 300 [200–350] mcg in scalp block vs. non-scalp-block groups, respectively; p = 0.0384).

Table 1 Characteristics of patients who received scalp block and those who did not receive scalp block.
Full size table

Scalp block and IDH1 mutation are associated with improved PFS but not OS

Scalp block was associated with longer PFS (Fig. 2A, median PFS: 15.17 [8.87–37.37] vs. 10.77 [6.97–12.77] months in scalp block vs. non-scalp-block groups, respectively; p = 0.0018). However, scalp block was not associated with longer OS (Fig. 2B, median OS: 43.70 [22.50–61.13] vs. 31.47 [18.83–36.83] months in scalp block vs. non-scalp-block groups, respectively; p = 0.4929). Although the IDH1 mutation was associated with longer PFS (Fig. 3A, median PFS: 37.37 [8.63–59.53] vs. 10.90 [8.27–13.77] months in IDH1 mutation carrier vs. noncarrier groups, respectively; p = 0.0149), the IDH1 mutation was not associated with longer OS (Fig. 3B, p = 0.4021).

Figure 2
figure 2

Oncological outcomes between patients who received scalp block and those who did not.

Full size image
Figure 3
figure 3

Oncological outcomes between patients with and without IDH1 mutation.

Full size image

Higher pathology grade and tumor location are additional risk factors associated with reduced OS

Multivariate Cox regression analysis revealed that scalp block (hazard ratio [HR]: 0.436, 95% CI: 0.236–0.807, p = 0.0082), gross total resection (HR: 0.405, 95% CI: 0.227–0.721, p = 0.0021), and IDH1 mutation (HR: 0.304, 95% CI: 0.118–0.784, p = 0.0138) were associated with longer PFS. World Health Organization grade IV glioma (HR: 2.363, 95% CI: 1.117–4.999, p = 0.0246) was a risk factor for less favorable PFS than grade III glioma (Table 2). Risk factor analysis for OS revealed that a pathology grading of IV (HR: 4.256, 95% CI: 1.487–12.182, p = 0.0070) and infratentorial tumor location (HR: 11.038, 95% CI: 1.309–93.083, p = 0.0273) were associated with reduced OS, whereas gross total resection (HR: 0.258, 95% CI: 0.120–0.554, p = 0.0005) and adjuvant radiotherapy (HR: 0.128, 95% CI: 0.031–0.524, p = 0.0043) were associated with improved OS. Neither scalp block (HR: 0.492, 95% CI: 0.222–1.089, p = 0.0800) nor IDH1 mutation (HR: 0.728, 95% CI: 0.225–2.357, p = 0.5965) were associated with OS (Table 3).

Table 2 Risk factors of worse progression-free survival.
Full size table
Table 3 Risk factors of worse overall survival.
Full size table

The parallel analyses without overlapping included 39 patients and the results were summarized in the supplementary file (Supplementary Table 1 and Supplementary Table 2). In brief, the parallel multivariate Cox regression analysis also revealed that scalp block was associated with longer PFS. (HR: 0.095, 95% CI: 0.018–0.510, p = 0.0060; Supplementary Table 2).

Discussion

The major discovery of this study was that scalp block is an independent prognostic factor for recurrence profile in patients with high-grade glioma, regardless of differences in IDH1R132H mutation profile.

To the best of our knowledge, the major difference between previous reports and the present study is that presence of the IDH1R132H mutation was for the first time included in an analysis of the influence of anesthetic technique on high-grade glioma oncological outcomes. A multigroup collaboration in 2008 sequenced 20,661 genes in 22 glioblastoma multiforme tumor samples and identified a common point mutation in the metabolic gene IDH1 (R132H) that was present in 12% of patients with glioblastoma multiforme19. The proportion of IDH1R132H in the present study (15.2%) is comparable to that reported previously. The study results are also compatible with those of previous studies indicating that patients with glioma carrying the IDH1R132H mutation were younger and had longer PFS12,20. IDH1 is a key enzyme in Krebs cycle functions dependent on NADP+. Wild-type IDH1 messenger RNA and protein are commonly overexpressed in primary glioblastomas, which indicates that IDH1 activity is important to metabolic adaptation of high-grade gliomas21. In addition, IDH1 regulates hypoxia-inducible factors related to tumor angiogenesis and invasiveness22. This is relevant to glioma recurrence because glioma cells exhibiting invasive characteristics after resection surgery are more highly competitive for nutrients than are normal neuronal tissue. The beneficial outcomes observed among IDH1R132H mutant gliomas may occur through several mechanisms. First, IDH1R132H mutations result in neomorphic enzyme activity, catalyzing the NADPH-dependent reduction of α-ketoglutarate to R(-)-2-hydroxyglutarate23, which increases DNA methylation of the glioma. Data from a large cohort of 1,122 diffuse grade II–IV gliomas revealed that IDH-mutant gliomas with high levels of DNA methylation had more favorable clinical outcomes than those with low levels24. Second, IDH1R132H mutations inhibit glial cell proliferation through inhibition of Bcl-xL, which induces more apoptosis than wild-type IDH25. Third, IDH1R132H mitigates expression of hypoxia-induced factor-1α, which interferes with glioma angiogenesis22. Although the present study indicated improved PFS among patients carrying IDH1R132H, improved OS was not observed. This may be because the survival of patients with high-grade glioma has markedly improved in the last decade owing to a rapid increase in the use of adjuvant and concomitant chemotherapy and radiotherapy, earlier diagnosis, and advances in multimodal health care26. In addition, survival is influenced by more complex factors such as age, baseline performance status27, long-term care capability28, and gross total resection rate29. Therefore, the influence of IDH1R132H on OS may have been obscured in our study.

The potential mechanisms of the protective effects of scalp blocks against high grade glioma recurrence are organized into three aspects, namely the prevention of a surgical stress response, the systemic effects of local anesthetic, and the sparing effects of general anesthetic. Accumulating evidence reveals that surgical stress and the subsequent neuroendocrine and inflammatory responses may negatively effects tumor recurrence after curative surgery1. First, the surgical stress response activates the sympathetic nervous system and the hypothalamic–pituitary–adrenal axis, thus promoting tumor-associated angiogenesis30,31. Second, surgical stress increases catecholamine circulation, which may impair immune function through mechanisms including diminished cytotoxicity of natural killer (NK) cells32, reduced dendritic cell maturity, and a decreased Th1/Th2 ratio, which suppresses antimetastatic cell-mediated immunity33. Furthermore, glioma cells express beta-adrenergic receptors, and catecholamine may thus interact with the glioblastoma proliferation34. Because scalp blocks effectively reduce surgical neuroendocrine stress, with concomitant reductions in plasma cortisol and catecholamine levels during craniotomy35, the aforementioned negative effects may be attenuated. In addition to the attenuation effects against stress responses, local anesthetics may prompt systemic protective effects against glioma proliferation through various mechanisms. First, local anesthetic blocks N-methyl-d-aspartate-type glutamate receptors from mediating rapid excitatory neurotransmission in the central nervous system36,37. Extracellular glutamate in the cerebral cortex negates IDH1R132H suppression of gliomagenesis, thereby driving anchorage-independent growth and glioma progression38. Accordingly, negative effects of glutamate may be inhibited by the local anesthetics in scalp blocks. Second, transient receptor potential (TRP) channels regulate cell proliferation and death39. Local anesthetic (e.g., lidocaine) upregulates the TRPV1 channels and suppresses the TRPM7 channels, and both of these mechanisms protect against glioma cell proliferation40,41. Third, local anesthetic was reported to induce protective autophagy in a rat C6 glioma cell line42. Moreover, it weakened ZDHHC15 transcripts and reduced GP130 palmitoylation levels and membrane localization, thus impairing the growth and self-renewal of glioblastoma stem cells43.

General anesthetic agents may elicit multiple negative effects during the surgical resection of glioma. First, inhalational anesthetics were reported to profoundly suppress the effects of NK cells. A clinical study reported that sevoflurane induced a significant decrease in NK cells during cranial surgery44. Sevoflurane was also reported to attenuate NK cell–mediated cytotoxicity by inhibiting adhesion molecule leukocyte function-associated antigen-1 in an experimental study45. The importance of NK cells for glioma outcomes has been recently addressed; the IDH1R132 mutation promotes recruitment of NK cells in the brain and is associated with improved glioma prognosis46. In addition, novel therapies based on augmentation of NK cell recruitment and function are emerging in neuro-oncology47. Therefore, the negative effects of inhalational anesthetics on NK cells may lead to an immunosuppressant tumor microenvironment in patients undergoing glioma resection surgery. Second, the inhalational anesthetics may induce the negative gene expression effects of human glioma cells. Isoflurane and sevoflurane were reported to increase the migration, invasion, and colony-forming ability of human glioblastoma cells in vitro, and their tumor volume and invasion/migration were increased in vivo through increases in the expression of cell surface protein 4448. Third, both inhalational anesthesia and propofol were reported to increase the amount of circulating extracellular glutamate in patients undergoing neurosurgery49, which may promote glioma proliferation. Therefore, scalp blocks may attenuate the aforementioned negative effects through strong anesthetic-sparing effects50.

Intravenous anesthesia has also been proposed to result in less intraoperative immunosuppression than inhalational anesthesia during surgery1, and a recent meta-analysis revealed beneficial oncological outcomes of intravenous anesthesia in patients with nonglioma cancer. However, the current literature on glioma is limited and reveals nonsignificant effects18,51,52. In the present study, we observed comparably nonsignificant effects of intravenous anesthesia on oncological outcomes of patients with high-grade glioma. This may have been because of the unique characteristics of the glioma immune microenvironment. First, glioma cells may express cell surface proteins to sequester T cells in the brain53. Second, the blood–brain barrier acts as a physical barrier against the exchange of immune cells from extraneural tissues and the brain54. Third, natural killer cells are the least abundant immune cell infiltrating glioma cells (representing only 2.11% of all infiltrating immune cells55; and glioma cells secrete soluble immunosuppressive factors to suppress natural killer cell activity56. Because the protective immune effects of intravenous anesthesia are mainly exerted in the peripheral circulation, particularly through the activation of natural killer cells57,58, intravenous anesthesia may be less influential on the oncological outcomes of patients with glioma.

There are several limitations in this study. First, the study is single-centered, retrospective in nature, and with a relatively small sample. Second, despite the IDH1R132H mutation being the most influential genetic factor for postoperative oncological outcomes of gliomas, information concerning other potential factors such as the chromosomal 1p/19q codeletion, IDH2 or MGMT mutation, and ATRX mutation was not considered. Third, the number of patients with infratentorial tumors was much lower than those with supratentorial tumors. Hence, caution is needed when extrapolating the results to patients with infratentorial tumors. Fourth, steroids are not administered for patients with primary glioma in our institute; thus, the present results should be interpreted cautiously for patients receiving perioperative steroids.

In conclusion, our results indicate that scalp block is associated with favorable recurrence profiles in patients with high-grade glioma. However, given the limitations of this study and the complexity of the pathophysiology of glioma recurrence, our results should be considered exploratory rather than as a basis for practice change.