Patients with brain tumours, both primary and metastatic, are at an increased risk of experiencing seizures. The incidence of seizures amongst patients diagnosed with a brain tumour varies depending on certain factors, including tumour type and location. In a series of 1028 patients diagnosed with primary brain tumours, those with low-grade gliomas had a higher prevalence of epilepsy (85%) compared with patients with a diagnosis of glioblastoma (49%) [1] . Seizures are less common in metastases, reported in 24% in one series [2] . However, this is also dependent on the type of metastasis, with melanoma seemingly the most epileptogenic (67%), reflective of a tendency to haemorrhage and to involve the cerebral cortex.

Epilepsy can be a significant burden on the patient due to the physical and psychosocial effects of both the condition and the treatment required. Although there is strong evidence against the use of AEDs for seizure prophylaxis in patients with brain tumours, they are certainly still prescribed in this clinical context. Overall, the adverse effects of AED therapy, including side effects and drug interactions, outweigh the benefits gained in preventing seizures in this patient group.

Data on physician practice in Rhode Island demonstrated that prophylactic anticonvulsant administration was 55% in a cohort of 115 [3] . There were differences between specialties with 33% of radiation oncologists prescribing, compared with 81% of neurosurgeons. There are limited data on current practice in Australasia and so here we aim to establish the prescribing patterns amongst Australasian neurosurgeons and to determine if there are some particular factors that influence the decision to be a prescriber.

A total of 91 responses were received from the 141 NSA members to whom the survey was distributed (64.5% response rate). Of these, 80 and 11 were neurosurgeons currently practising in Australia and New Zealand (NZ), respectively. The majority of the Australian responders were from New South Wales (NSW) [n = 26], Victoria (VIC) [n = 23] and Queensland (QLD) [n = 18], with fewer contributions from the remaining states or territories. No responses were received from Tasmanian (TAS) neurosurgeons. The Northern Territory (NT) was not included in the analysis as there is no neurosurgical unit based there. Experience was stratified by decade of admission to the NSA, noting the earliest respondent practising since 1971.

Seizures can be a major burden on a patient with potentially significant quality of life and lifestyle implications. In patients with a brain tumour who have suffered a seizure/s, there is no controversy regarding the use of AEDs to prevent future events. However, the prophylactic use of AEDs in patients with a newly diagnosed brain tumour, who have not had a seizure, remains an issue of debate and practice varies. This may reflect lack of clarity in the available literature.

A meta-analysis of four randomised controlled trials (RCTs) and eight observational studies, found that prophylactic use of AEDs (specifically phenytoin, phenobarbital and divalproex sodium) is not effective in preventing first-onset seizures nor in improving the seizure-free and overall survival in patients with brain tumours [3] . In another meta-analysis of five RCTs that evaluated the efficacy of AED prophylaxis vs. no treatment or placebo to prevent seizures in patients with brain tumours who had no history of epilepsy, the conclusion was that there was no evidence to support their use regardless of neoplastic type [4] .

Given that these patients will often proceed to a craniotomy, it is also important to consider the potential for seizures in the perioperative period. The question of perioperative seizure prophylaxis in patients with intrinsic brain tumours was explored in a RCT with a cohort of 123 patients, 62 of which were randomised to receive a 7- day postoperative phenytoin course (prophylaxis group) and 61 receiving no prophylaxis (observation group) [5] . They found that the incidence of early postoperative seizures (defined as seizures within 30 days of surgery) was 8% in the observation group compared with 10% in the prophylaxis group (p = 1.0). The incidence of clinically significant early seizures was 3% and 2% in the observation and prophylaxis groups, respectively (p = 0.62). In addition the prophylaxis group experienced significantly more adverse events (18% vs. 0%, p < 0.01). The study concluded that the low baseline rate of perioperative seizures in patients with brain tumours raises concerns about the routine use of prophylactic phenytoin in this patient population.

Results of this survey indicate that prophylactic AED prescribing in this clinical context is not uncommon in Australasia despite the evidence, with a 25.3% rate of prescription. It was interesting to find that of the 11 respondents from NZ, none of them prescribe prophylactic AEDs. Whether or not this is due to specific guidelines in NZ is uncertain. Practising neurosurgeons in WA tended towards prescribing (75.0%), noting that there were only 4 respondents from WA. Neurosurgeons in SA and ACT were less inclined to prescribe prophylactic AEDs. It was also interesting to note that there weren’t any major differences in prescription rates based on neurosurgeon experience. The obvious exception was the decade between 1981 and 1990 where there was a 50.0% prescription rate, although there were only 6 respondents from this era.

The main concern regarding prescription of AEDs relates to their potential adverse effects including side effects and drug interactions. AEDs can be associated with cognitive side effects, which impair patients who are already often cognitively affected by their condition and associated adjuvant treatments [6] . Although phenytoin has traditionally been the AED of choice amongst neurosurgeons, including in this study, it is a drug not without potential hazards. In one study it was found that patients on phenytoin, who are undergoing radiotherapy and weaning steroids, a common clinical scenario in patients with brain tumours, have an increased risk of erythema multiforme and Stevens-Johnson syndrome [7] . Phenytoin has a narrow therapeutic index and thus regular monitoring is required to avoid toxicity.

Levetiracetam is becoming a popular alternative AED drug of choice given that it is generally well tolerated by patients. In addition, unlike phenytoin, monitoring of plasma levetiracetam levels is not necessary and it is therefore easier to administer with dose adjustments made based on clinical response. Interestingly, presently under the general schedule of the Australian Government Department of Health’s Pharmaceutical Benefits Scheme (PBS), levetiracetam is authorised only for the treatment of partial epileptic seizures which are not controlled satisfactorily by other anti-epileptic drugs [8] . This clearly precludes it from use as a prophylactic medication, unless prescribed as a non-PBS medication, which has potential cost implications for patients. Although seemingly safer than phenytoin, levetiracetam can cause side effects also with mood disturbance/irritability reported in approximately 19% of patients taking the drug [9] .

Aside from side effects, the other major concern with the use of AEDs is the potential for drug-drug interactions. Through induction of the cytochrome P450 enzyme system, phenytoin and other prescribed AEDs can alter the metabolism of chemotherapeutic agents and steroids, reducing their effectiveness [10] . Conversely, the chemotherapeutic agents can induce metabolising enzymes, altering serum levels of AEDs [10] .

Here we found that of those who prescribe, 68.2% choose phenytoin as a first-line prophylactic agent. Levetiracetam was the next most commonly prescribed first-line agent (22.7%). Interestingly although sodium valproate can be associated with thrombocytopenia and it impairs platelet aggregation and thus can potentiate surgical bleeding [11] , it was still prescribed by 9.1% of respondents as a first-line agent. On the other hand, there is some suggestion that patients with a diagnosis of glioblastoma who take sodium valproate, have a significantly longer survival when compared with patients not on AEDs or treated with other AEDs [12] . The proposed mechanism is related to its activity as a histone deacetylase inhibitor (HDI), with potential antitumour effect.

Through this simple survey, we have been able to explore the general prescribing patterns amongst Australasian neurosurgeons with regards to prophylactic AEDs. However, there are some limitations of this study that warrant discussion. Firstly, although there was a good overall response rate (64.5%), without answers from the entire target group, there is obvious potential for bias. Also neurosurgeons who are not NSA members did not receive the invitation email. Amongst prescribers, there were several non-responders to the remaining questions. One of the questions was omitted completely from the final analysis due to a poor response rate. Furthermore, the style of the survey questions were closed and as such, further analysis is difficult. For example, although tumour location was selected as an important influencing factor by 95.0% of prescribers, there was no field to specify which particular locations are important. Overall however, the study was designed to gauge general prescribing patterns amongst Australasian neurosurgeons and to hopefully stimulate further debate regarding this topic.

Although the available evidence does not support the use of AEDs for preventing seizures in patients with brain tumours, clearly it is still commonplace amongst practising neurosurgeons, at least in Australia. Phenytoin remains the drug of choice, although levetiracetam is gaining popularity given its favourable side-effect profile and easier management. In terms of duration, 6 weeks was most commonly reported. The tumour location and a history of previous seizure(s) were the most important factors in the decision of prescribers. It will be interesting to see if current trends continue, and a follow-up survey will be worthwhile in the future.

The authors declare that they have no financial or other conflicts of interest in relation to this research and its publication.