Forty seven patients with PA and 50 surgically treated NFPAs were included. The mean age at presentation was 54 ± 15 years (range 23–88 years) for the PA group versus 61 ± 14 years (range 27–88 years) for the NFPA group. The PA group was significantly younger (independent t test, t = 2.25, p = 0.03). There was a male predominance in both the PA and NFPA groups, with a male:female ratio of 33:14 and 32:18, respectively. There was no significant difference in gender distribution between the groups (Fisher’s Exact, p = 0.53).
Clinical features of PA group
Clinical features of the PA group are shown in Table 1. Precipitating factors were identified in 15 cases:11 cases of existing hypertension and 4 cases of antiplatelet/anticoagulant use. Headache was the commonest presenting symptom in 42/47 cases (89%).
The mean PAS was similar between the conservatively and surgically managed groups (independent t test, p = 0.14; Table 1). The proportion of these groups with a PAS ≥ 4 was also similar between the conservatively and surgically managed groups (9.1 vs. 25.0%, Fisher’s Exact, p = 0.41; Table 1). The severity of apoplexy grade was similarly distributed between the groups (Chi-squared = 1.92, p = 0.38; Table 1).
Management of PA group
Most patients with PA were treated with trans-sphenoidal surgery (36/47, 77%) and a smaller proportion were managed non surgically (11/47, 23%). Two patients in the conservatively managed group of PA patients eventually had surgery at 10 and 23 months for significant residual adenoma that was non-functioning in one patient and growth hormone secreting in another. Given that the presentation with apoplexy was managed conservatively, these patients remained in the conservatively managed group. In the surgical group, the median time to surgery after presentation was 5 days (range 1–167 days).
Five (14%) patients with PA had postoperative complications, including 1 case each of post-op CSF leak (managed with a lumbar drain), hyponatraemia, depression, hospital acquired pneumonia and atrial fibrillation. Three patients in the surgically managed PA group underwent further surgery for residual tumours and 8 received radiotherapy.
The majority (45/47, 96%) of patients with PA required hormone replacement in the form of: steroids (38/47, 81%), thyroxine (38/47, 81%), sex hormones (28/47, 60%), growth hormone (25/47, 53%) and desmopressin (3/47, 6%) at their last clinical review.
The median time to first MRI scan was 7 days for the PA group (range 1–224 days). This time interval was greater than 4 weeks in 5 patients who presented in a delayed fashion to the local endocrine team. Two out of the 5 patients also had complex medical/biochemical disorders that required management prior to the first MRI scan. Once admitted to the neurosurgical centre, the median time to MRI scan was 1 day (range 1–19 days).
Imaging characteristics are shown in Table 2. The majority of PA and NFPA patients had a sellar variant of sphenoid sinus, with no significant difference in the distribution of sinus types between the groups (Chi-squared = 1.73, p = 0.42; Table 2). A greater degree of sinus opacification was present in patients with PA (Chi-squared = 22, p < 0.001; Table 2).
The median (range) sphenoid sinus mucosal thickness was 2.0 mm (0.5–6.0 mm) in the PA and 0.5 mm (0.5–2.0 mm) in NFPA groups, respectively (Mann–Whitney, p < 0.001). Sphenoid sinus mucosal thickness greater than 1 mm was noted in a higher proportion of PA (29/47; 62%) than NFPA group (3/50; 6%) (Chi square = 35, p < 0.001; Table 2). Sphenoid sinus mucosal thickness greater than 3 mm was only apparent in the PA, and not NFPA, group (23 vs. 0%; Table 2). The median (range) sphenoid sinus mucosal thickness was similar between the conservative −2.0 mm (0.5–4.0 mm), and surgically managed −2.0 mm (0.5–6.0 mm), PA patients (Mann–Whitney, p = 0.53).
Case examples of SSMT in NFPA and PA patients are shown in Fig. 2. Figure 3 demonstrates a case where imaging was available pre and post-apoplexy, demonstrating the development of SSMT shortly after the symptoms of PA became apparent.
Multivariate analysis of both PA and NFPA cases revealed PA to be the only factor associated with SSMT (OR 0.043, 95% confidence interval = 0.012–0.16; p < 0.001) (Table 3).
Multivariate analysis of the PA group alone, revealed that the time to scan (OR 0.12, 95% confidence interval = 0.026–0.54; p = 0.006) and severity of apoplexy grade (OR 7.29, 95% confidence interval = 1.10–48.40; p = 0.04), were the most important factors associated with SSMT (Table 4).