**1. Introduction**

Cerebral radiation necrosis is a frequent complication of current treatment algorithms for malignant brain tumours [1]. Radiotherapy is an integral part of first-line therapy for primary brain tumours like malignant gliomas as well as brain metastases [2,3]. Because the majority of malignant brain tumours are incurable, recurrent disease is almost always inevitable and a second course of radiotherapy can be considered under some circumstances [4], which further increases the risk of cerebral radiation necrosis. Pathophysiologically, cerebral radiation necrosis is characterized by capillary collapse and liquefaction necrosis of brain tissue, which causes an inflammation and vascular endothelial growth factor (VEGF) A-mediated disruption of the blood–brain barrier (BBB) [1,5,6]. Local inflammation can additionally cause necrotic areas to spread and the associated brain edema can greatly exceed the area of BBB disruption. Therefore, cerebral radiation necrosis can cause significant morbidity.

Established therapy for cerebral radiation necrosis is the administration of high-dose corticosteroids [7]. The most commonly employed dexamethasone is a high-potency, long-acting corticosteroid with a biological half-life of 36 to 54 h, which causes profound suppression of the hypothalamus–pituitary–adrenal hormone axis [8]. Patients, especially those on long-term treatment, frequently experience several side effects including weight gain, body edema, skin thinning, striae rubrae, proximal myopathy, steroid-induced diabetes, sleep disturbance, mood changes and sometimes steroid psychosis or depression, osteoporosis, thrombosis, and infections [9]. Bevacizumab is an antibody targeting VEGF-A as a mediator of angiogenesis [10] and established targeted therapeutic approach in some cancer entities including breast and colorectal cancer [11,12]. Much hope was therefore placed in its possible efficacy in glioblastoma (GB). While the first phase II trial of bevacizumab and irinotecan in recurrent glioblastoma with dramatic improvement in MRI presentation (at least a partial response in 63% of patients) sparked enthusiasm [10], three subsequent phase III trials of first-line therapy failed to show any prolongation of overall survival [13–15]. However, similar MRI improvements with reduced gadolinium contrast enhancement had been observed in these studies [13–15], revealing the ability of bevacizumab to reduce the permeability of the BBB without a significant anti-GB effect. As a consequence of the tightening of the BBB, bevacizumab also allowed reducing the corticosteroid doses reported, e.g., in the AVAglio trial (BO21990) [15]. This effect of bevacizumab has been used clinically in small patient collectives as a treatment option for cerebral radiation necrosis [16–18]. However, bevacizumab has not been approved by the European Medicines Agency (EMA) for this indication. Nevertheless, when dexamethasone has to be discontinued as a treatment for patients with cerebral radiation necrosis due to adverse or insufficient antiedematous effects, bevacizumab is an option as part of an individual, off-label therapeutic approach that frequently allows the tapering off of parallel dexamethasone. Since such patients have commonly been treated with dexamethasone for weeks or months, consecutive adrenal insufficiency (AI) has to be considered. The clinical symptoms of AI are nonspecific, and symptoms like lethargy, weakness, and nausea can be misinterpreted as consequences of the tumour treatment or the tumour itself. It may also be challenging to differentiate between AI occurring as a consequence of terminated dexamethasone treatment, which should be substituted with hydrocortisone, and the recurrence of cerebral edema, which is best treated with dexamethasone. In order to evaluate the frequency of AI in brain tumour patients treated with dexamethasone, we chose a collective of bevacizumab-treated patients because corticosteroids can often be terminated and thus basal cortisol can be analysed accurately in this patient collective.
