Predictive Factors of Postoperative Peritumoral Brain Edema after Meningioma Resection
Correspondence Address: Source of Support: None, Conflict of Interest: None DOI: 10.4103/0028-3886.333500
Source of Support: None, Conflict of Interest: None
Keywords: Brain invasion, cerebral edema, complication, meningioma, seizure
Meningioma is one of the most common primary tumors in the central nervous system. Complete resection can be curative for benign meningiomas (WHO I grade). Patients often presented with minor clinical symptoms; thus, a postoperative complication can be devastating and dramatically worsen the outcome. Postoperative Peritumoral brain edema (PTBE) is not an uncommon complication following meningioma resection. A letter recently published by Saffarian et al. proposed the term “wound meningioma syndrome.” The syndrome was defined as an exacerbation of the edema in the brain surrounding malignant meningiomas. The mechanism underlying postoperative PTBE is still under debate. This raises the interest to identify its risk factors in meningiomas.
A retrospective survey was conducted on patients who underwent meningioma resection from June 2008 to June 2018. Exclusion criteria were: 1) multiple intracranial tumors; 2) prior cerebrovascular disease or surgery (e.g., tumor embolization); 3) patient with abnormal coagulation function. Patients routinely received head computed tomography (CT) in the postoperative 24 h. The second head CT was planned within one week based on patients' clinical manifestations. All Patients were followed up for at least 30 postoperative days. Patients with missing clinical data were excluded.
Age, sex, chief complaint, duration of symptom and tumor localization, shape (regular/irregular), T2 signal (hypo-, iso-/hyper-intensity), presence of preoperative PTBE and tumor maximal diameter (mm), tumor volume (cm3), midline shift, preoperative mannitol use and tumor consistency (soft/firm), superior sagittal sinus (SSS) violation, surgical Simpson grade, operation time and intraoperative blood loss, antiepileptic drugs use and hospital stay were recorded. All pathological samples were reviewed by Dr. Chen YZ according to the 2016 WHO Classification of Central Nervous System Tumors' criteria. Seizure types were classified according to International League Against Epilepsy 2017 version. Edema index (EI) was calculated by dividing preoperative PTBE volume plus tumor volume by tumor volume in T2-weight magnetic resonance imaging (MRI). “None” (EI = 1.0), “moderate” (EI = 1.0–2.0), and “severe” (EI > 2.0) edema were defined accordingly. Preoperative mannitol use beyond 3 consecutive days was counted. Tumor consistency was assessed by two surgeons (Dr. Lin SH or Dr. Huang HT) during the operation. A tumor is defined as “soft” if it can be partially resected by blunt dissection with suction and bipolar coagulation. Postoperative PTBE was defined if postoperative head CT showed there was a progressively exacerbating PTBE (increase in volume compared to preoperative PTBE) after meningioma resection, with or without hemorrhage [Figure 1]. Clinical outcome of patients having postoperative PTBE was assessed by the Glasgow Outcome Scale (GOS) on the 30th postoperative day.
Data were analyzed with SPSS 19.0 (Chicago, IL, USA). Statistical analyses of categorical variables were carried out using Chi-square tests or Fisher's exact tests as appropriate. Statistics of means were carried out with independent Student's t-tests or non-parametric tests. Correlations of risk factors and postoperative PTBE were evaluated by binary logistic regression. The results were demonstrated as odds ratio (OR) ± confidence interval (CI). The interaction of variables was assessed according to Andersson et al. using the relative excess risk due to interaction (RERI), the attributable proportion due to interaction (AP), and the synergy index (S). A P value <0.05 was considered significant. All reported P values are 2-sided.
A series of 117 consecutive patients at the department of neurosurgery were eligible for the study. There were 88 females and 29 males. Age ranged from 26 to 79 years old (52.1 ± 11.7). The overall postoperative PTBE rate was 11.1% (13/117). Patients' information and clinical characteristics appear in [Table 1] and [Table 2]. For continuous variables, values are expressed as mean ± standard deviation. For dichotomous variables, values are presented as the percentage (number).
Eight patients (6.8%) presented with a preoperative seizure. Among them, four had a focal aware myoclonic onset, two had a generalized tonic-clonic onset, one had a generalized tonic onset and one had an absence onset.
For 13 patients with postoperative PTBE, the symptoms mostly started within the postoperative 72 h. However, two patients presented with delayed symptoms occurring on the 5th and 11th postoperative days. Two patients (15.4%) developed a depressed level of consciousness and were received secondary decompressive surgery. Unilateral limb weakness (eight cases, 72.7%), headache (three cases, 27.3%), unilateral limb numbness (one case, 9.1%), and slurred speech (one case, 9.1%) were presented in other 11 conservatively treated patients (84.6%). GOS in the 30th postoperative day were 5 in 4 patients (30.8%), 4 in 7 patients (53.8%), and 3 in 2 patients (15.4%).
Univariate and multivariate analysis
Univariate analysis showed that the complication rate of postoperative PTBE between patients having a seizure as the chief complaint and those not having was significantly different (P = 0.044). Tumors had moderate preoperative PTBE (P = 0.100) and located on the parietal lobe (P = 0.078) were prone to develop the complication. No statistically significant differences were found between patients having and not having postoperative PTBE in tumor volume, SSS violation, Simpson grade, or WHO classification [Table 2] and [Table 3].
Binary logistic analysis of risk factors showed that “preoperative seizure” and “prominent nucleoli” were the significant independent risk factors for postoperative PTBE. The variate “parietal localization” was closely related to the complication but it did not stand up to the statistical test [Table 4]. The variate “firm tumor consistency” was observed to be a suppressor for the variates “preoperative seizure,” “prominent nucleoli,” and “parietal localization.”
There was no biological interaction between preoperative seizure and parietal localization for postoperative PTBE (RERP 0.375, ranged from −25.192 to 25.942, AP 0.038, ranged from −2.474 to 2.549, S 1.043, ranged from 0.057 to 19.073).
Postoperative PTBE was also referred to as “postoperative hemorrhagic infarction,” “venous edema,” or “brain swelling” in the literature.,,,,, Clinical symptoms of postoperative PTBE are variable, including symptoms of high ICP and localized symptoms caused by eloquent cortex impairment. The overall outcome was difficult to be assessed because the definitions for the complication varied among the previous studies.
Preoperative seizure increased postoperative PTBE
The preoperative seizure occurs in 10–50% of patients with meningioma. There were very few papers that studied the association of preoperative seizure and postoperative complications in meningiomas, and most studies focus on the relationship between preoperative and postoperative seizure.,,, Chow et al. reported that preoperative seizure predicting the occurrence of a postoperative seizure, and postoperative seizure increases the risk of cerebral edema and hemorrhage in the first postoperative week. In our study, we found that preoperative seizure was the independent risk factor for postoperative PTBE after meningiomas resection.
35.9% of the patients received preoperative antiepileptic drugs (sodium valproate or oxcarbazepine), and 67.5% of the patients received them postoperatively. There was no significant association between antiepileptic drugs application and postoperative PTBE (P = 0.261 for preoperative use and P = 0.861 for postoperative use, respectively).
62.5% of the patients with preoperative seizures had meningioma on the parietal lobe (P = 0.103). Although seizure occurrence was associated with the tumor location, the negative interaction between preoperative seizure and parietal localization suggested that preoperative seizure may induce postoperative PTBE through a separating mechanism.
Studies had revealed that preoperative seizure increased in patients having meningiomas with brain invasion,, or preoperative PTBE,,, which may contribute to postoperative PTBE. But our study did not find these association (P = 1.000 for brain invasion and P = 0.283 for preoperative PTBE).
The predisposing effect of preoperative seizure on postoperative PTBE remained unclear. While excluding the above mechanisms, it is of great possibility that this effect is related to the neurotransmitters or inflammatory mediators from the tumor. Inflammatory mediators are involved in tumorigenesis and many chronic central nervous system conditions including seizure., Removal of the tumor abruptly withdraws the surrounding brain tissue from vasodilators such as prostaglandin E2, which could lead to postoperative PTBE. This theory needs further investigation.
Brain invasion in preoperative and postoperative PTBE
It is reported that preoperative PTBE is more prominent in brain invasive meningiomas than non-invasive ones.,, But studies focus on postoperative PTBE are limited. Our results agreed on that brain invasive meningiomas had a significantly higher EI (3.0 ± 2.2 versus 1.8 ± 1.7, P = 0.001). However, we did not find a relationship between brain invasion and postoperative PTBE (P = 0.604).
We observed that meningiomas with moderate preoperative PTBE (EI = 1.0–2.0) had a higher chance of postoperative PTBE (21.4% versus 7.9%, P = 0.100), demonstrating that preoperative and postoperative PTBE was not simply linear correlated. An anatomical study discovered that meningiomas with preoperative PTBE had more vessels crossing the brain-tumor interface. Damage of the surgical cleavage plane between the tumor and cerebral cortex may lead to postoperative venous infarction. But For tumors with a severe PTBE (EI >2.0), those regions received neither the blood supply from the internal nor external arteries. During the operation, we usually observe a poor vascularization in these regions of PTBE, which may be insensitive to the hemodynamic change after the tumor resection.
Venous injury and postoperative PTBE
Venous injury is the most well-known immediate cause for venous edema or infarction. Injury of cortical veins or dural sinuses may cause local or distant venous congestion, which gives rise to postoperative PTBE. The incidence of postoperative venous infarction in meningiomas is 2.0–4.0%.,, Studies showed that convexity and parasagittal meningiomas had the highest complication rate. This may be attributed to the damage of a seemingly unimportant cortical vein among the complex venous structure in the midline.,
We found that meningiomas located on the parietal lobe were more prone to postoperative PTBE. The predilection site of postoperative PTBE and venous infarction was overlapped, indicating venous injury is a major source of postoperative PTBE.
However, postoperative PTBE did not necessarily mean venous infarction. While both lesions present as a newly developed hypodense lesion in postoperative CT, MRI is the best diagnostic modality to distinguish these two situations. In our study, postoperative MRI in two patients having postoperative PTBE showed no evident infarction in the DWI sequence, suggesting that there were still other mechanisms behind. But for developing countries, many patients would skip postoperative MRIs for saving costs. Also, some patients may be delayed due to a long appointment time (average 7 days for inpatients in our hospital). Therefore, CT remains to be the most convenient and inexpensive method for postoperative review. CT helps distinguishing arterial and venous injuries according to the pattern of cerebral edema. Once a large arterial occlusion is excluded, the management of cerebral edema caused by postoperative PTBE or venous infarction is similar, including osmotic diuretic such as mannitol and steroid such as dexamethasone.
The meta-analysis revealed that venous infarction occurred in 2% of patients with meningioma invading SSS. The complication rate doubled if the SSS was “aggressively” resected and reconstructed. The postoperative PTBE rate of meningiomas invading SSS in our serial was 7.1% (1/14). The Chi-square test found no significant association between SSS and postoperative PTBE (P = 0.519). The logistic analysis showed that the variate “tumor with parietal lobe localization” was a better predictor than the variate “tumor originated from falx cerebri” for the criterion “postoperative PTBE.” For tumors originated from SSS, we aim to maximally preserve the structure of the SSS, and nearly 50% of patients received Simpson grade III-IV resection. This surgical attitude may lower the postoperative PTBE rate in SSS invasive meningiomas in the study.
Pathohistological prominent nucleoli increased postoperative PTBE
Interestingly, prominent nucleoli was the only significant histological feature associated with postoperative PTBE. Nucleolar features can predict prognosis in various types of cancer. Prominent nucleoli or macronucleoli indicated a rapid proliferation in the tumor cells. It was also reported that prominent nucleoli increased meningioma recurrence., In our study, tumors with prominent nucleoli had an obviously larger tumor volume (78.9 ± 79.4 cm3 versus 42.6 ± 60.5 cm3, P = 0.025) and EI (2.7 ± 2.3 versus 1.9 ± 1.7, P = 0.013). Nevertheless, we should notice that tumors with prominent nucleoli may induce the formation of PTBE though producing growth factors such as vascular endothelial growth factor, which may increase tumor vessel permeability and disrupted BBB.
Tumor volume was not related to postoperative PTBE
Studies pointed out that “normal perfusion pressure breakthrough (NPPB)” may play a role in postoperative PTBE in meningioma. It was assumed that the presence of a mass would also impair the autoregulation of blood pressure in the surrounding brain tissue. Removal of a mass, such as a huge and vascularized meningioma, may suddenly increase the local blood flow, causing postoperative PTBE and intracerebral hemorrhages.
However, neither tumor volume (P = 0.931) nor intraoperative blood loss (P = 0.340) was related to postoperative PTBE in our study. The total volume of tumor and preoperative PTBE was also insignificant for the complication (P = 0.757). So far there was no evidence to support this hypothesis.
The outcome of postoperative PTBE was relatively good. No patients died in our serial. Patients having postoperative PTBE had a significantly longer hospital stay (P = 0.002). While two patients required secondary surgery, the other 11 patients were conservatively treated.
Four patients (30.8%) recovered without obvious sequela (GOS of 5). Nine patients (69.2%) were discharged with certain neurological deficits (7 patients with GOS of 4, 2 patients with GOS of 3). All of them had a gradual recovery of limb strength to grade 3-4 in the 30 days follow-up.
Postoperative PTBE is not a rare complication in meningioma resection. Preoperative factors including preoperative seizure and tumor parietal localization predisposed patients to the complication. Careful venous protection during the operation may be helpful for patients with high risks. The complication may occur later than 1 week after the surgery in some cases. Therefore, prominent nucleoli showed in the postoperative pathology should alter the surgeons to the complication. Close observation may improve patients' clinical outcomes.
All procedures performed in studies involving human participants were in accordance with the ethical standards of the local review board (K2018-022) and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards.
Informed consent was obtained from all individual participants included in the study.
Research involving animal participants This article does not contain any studies with human participants or animals performed by any of the authors.
Financial support and sponsorship
Conflicts of interest
There are no conflicts of interest.
[Table 1], [Table 2], [Table 3], [Table 4]