We describe three patients representing a spectrum of pituitary GPA and variable degrees of systemic involvement. All three patients had disease features in keeping with the updated classification criteria for GPA recently endorsed by the American College of Rheumatology/European League Against Rheumatism (ACR/EULAR) 2022 [10]. There is no consensus regarding the diagnostic criteria to confirm pituitary involvement in GPA. GPA-related hypophysitis is commonly diagnosed in patients who present with pituitary-related hormonal dysfunction or visual defects in the presence of positivity for serological markers, in particular PR3-ANCA, with a history of GPA that can involve various organs including kidneys, lungs and sinuses. A large multicentre retrospective study of 819 patients with GPA reported 9 (1.1%) had pituitary involvement, 8 of which presented late after GPA initially manifested in other organs [11]. However, isolated pituitary GPA without systemic involvement has been reported [12]. A follow-up study of progression of GPA limited to the head and neck in 24 patients with an average follow-up of 6.8 years, showed that the disease spread to additional sites in more than half of the cohort (n = 14), but only 2 of 13 patients with disease initially limited to the head and neck developed pulmonary disease, and none developed renal disease [13]. Furthermore, patients with GPA-related hypophysitis may be negative for c-ANCA. 50–70% of patients with localised GPA may have negative ANCA testing, whereas 80–95% with systemic disease are positive for c-ANCA antibodies [10,11,12]. Pituitary biopsy does not routinely need to be undertaken if there is a clear clinical context of serological and systemic evidence of GPA. Confirmatory pituitary biopsy is required if the clinical scenario does not meet the criteria to diagnose GPA and if other aetiologies of a pituitary mass need to be ruled out [14,15,16]. Granulomatous inflammation and inflammatory infiltrates are typically found but are often non-specific in biopsy samples of pituitary tissue with GPA [14, 17].

Radiological findings of pituitary GPA have been described in case studies and case series and can be variable according to the central nervous system (CNS) structure(s) involved [3, 14]. Gadolinium-enhanced pituitary MRI is the optimum method for radiological assessment of the pituitary. Kapoor et al. reported that the radiological finding of a sellar mass with peripheral enhancement, central cystic changes and pituitary stalk compression were most commonly observed in patients with pituitary GPA. Other reports described radiological features of an enlarged gland with either heterogeneous or homogeneous enhancement [3, 15]. In addition, there is increased enhancement and thickening of the infundibulum, especially its superior portion [18]. Furthermore, a loss of the hyperintense signal in the posterior pituitary on T1-weighted sequences, explained by decreased AVP content in the posterior pituitary, was reported to be strongly correlated with the clinical manifestation of cranial AVPD in patients with GPA [3, 14, 15]. Importantly, normal MRI findings do not necessarily exclude pituitary involvement and anterior or posterior pituitary hormone dysfunction may be the only manifestation [4].

Given its rarity, there no studies of optimum immunosuppressive strategies in GPA-related hypophysitis, and specific recommendations are lacking in current international consensus guidelines (e.g., ACR, KDIGO, EULAR) [7, 19, 20]. Nor is this disease feature specifically considered within the spectrum of CNS or endocrine disease evaluated in commonly used disease assessment tools such as the Birmingham Vasculitis Activity Score (BVAS) [21]. Given the risk of pituitary failure, we suggest hypophysitis is regarded as a life-threatening disease complication warranting intensive treatment. We propose remission induction with glucocorticoids (prednisolone 60 mg daily followed by a weekly reduction to 45 mg, 30 mg, 25 mg and 20 mg and thereafter to be reduced by 5 mg every 2 weeks). We also propose to start RTX early initially 2 × 1 g doses 2 weeks apart and 1 g 6 monthly thereafter to achieve B-cell depletion.

In the published literature, the combination of cyclophosphamide and high-dose glucocorticoids has been used most frequently, and historically was shown to reduce 1-year mortality from 80% to 10–20% in patients with systemic GPA [9, 22]. However, toxicity of cyclophosphamide and the high relapse rate remain a challenge. The past decade has seen increasing use of RTX, a B lymphocyte-depleting monoclonal anti-CD20 antibody targeting B lymphocytes, including those producing ANCAs, both for remission induction and maintenance therapy. The landmark Rituximab in ANCA-Associated Vasculitis (RAVE) trial showed that treatment with RTX and glucocorticoids was equivalent to the standard combination of cyclophosphamide with glucocorticoids for attainment of disease remission at 6 months [23]. RTX is now recommended as a potential alternative to cyclophosphamide as first line therapy in current international consensus guidelines, although specific data on rare disease manifestations such as hypophysitis are lacking. Promisingly, a recent case series of 11 patients with CNS involvement of ANCA-associated vasculitis by Krishna et al. showed that RTX was as effective as cyclophosphamide in remission induction in patients with CNS involvement of GPA [24]. The use of cyclophosphamide and rituximab in combination has been studied in one randomised trial and several uncontrolled series, an approach that may facilitate rapid disease control and glucocorticoid-sparing, which was used successfully in one of our patients [25,26,27,28]. A regimen using glucocorticoid and RTX alone might be more suitable for patients where pituitary involvement is the predominant manifestation, but larger studies will be required to confirm our observational experience [29].

Regarding remission-maintenance, several studies have confirmed the superiority of regular rituximab infusions to conventional disease-modifying immunosuppressants, particularly in those with relapsing disease, administered either at fixed 4–6 monthly intervals or tailored to circulating B lymphocyte or ANCA levels, and this approach is now recommended depending on availability of rituximab [22, 24, 30]. Novel therapies for AAV, including GPA, include avacopan, a C5a-receptor antagonist that has shown efficacy as a glucocorticoid sparing agent, when used alongside rituximab or cyclophosphamide, in a recent large randomised controlled trial [31]. Avacopan was recently approved for treatment of AAV in the UK, although specific data regarding treatment of rare disease manifestations such as hypophysitis are lacking.

All our patients received combination therapies, including high-dose glucocorticoids with biological and/or immunosuppressant drugs (RTX, methotrexate, mycophenolate mofetil and cyclophosphamide), depending on the severity of pituitary GPA and the presence of other systemic manifestations (Table 2). In all cases, there was radiological evidence of GPA-related hypophysitis with excellent pituitary response to therapy.

All 3 patients continue to have cranial AVPD and remain on DDAVP supplementation. Interestingly, our observations differ from those of Kapoor et al., who found that AVPD resolved in 4 out of 6 patients [3]. In addition, patient 1 and 3 remain on hormone replacement for secondary hypoadrenalism and secondary hypothyroidism.

Patient 2, who had localised pituitary GPA and limited sinus involvement only, remains in full remission with 6-monthly RTX infusions and no longer requires glucocorticoid maintenance therapy. In contrast, patient 1 had GPA for 28 months before presenting with GPA-related pituitary hypophysitis and therefore required a longer treatment period to achieve clinical and radiological responses.

RTX treatment was generally well tolerated and none of our patient reported any significant infections. However, patients experienced significant weight gain while on high-dose glucocorticoid therapy.