Outcomes of endoscopic transsphenoidal surgery for Cushing's disease

With an estimated annual incidence of 1.7 per million [1], Cushing’s disease is rare. Untreated, it poses serious complications including osteoporosis, hypertension, dyslipidaemia, insulin resistance, and hypercoagulability [2] and is associated with a 4.8 fold increase in mortality rate [3‐5]. Patients who are in remission from CD have a mortality rate which decreases towards (although not reaching) that of the general population [6]. Endoscopic transsphenoidal surgery (ETSS) offers patients potential remission from Cushing’s disease, although long term surveillance is required as recurrence rates range from 5 to 22%% [7‐12].

Since the first report in 1997 [13], the selective removal of an adrenocorticotropic hormone (ACTH)-secreting pituitary adenoma by endoscopic transsphenoidal surgery has gained popularity as the first line treatment for Cushing’s disease. The primary goal of ETSS treatment in Cushing’s disease is to produce disease remission and to provide long-term control, while minimising complications. Remission rates are dependent on tumour size, preoperative MRI, cavernous sinus invasion, intraoperative visualisation of the tumour and pre- and postoperative ACTH and cortisol concentration [11]. Several studies also report pituitary neurosurgeon experience as a major factor for operative success [2, 14, 15].

Reported remission and recurrence rates after TSS for CD vary widely according to the criteria utilised to define remission [11], and in some studies due to limited patient numbers or short follow-up periods. Indeed, there is no clear consensus on how best to define post-operative remission; an early morning serum cortisol concentration < 138 nmol/L (5μg/dl) within 7 days of TSS is quoted in the 2015 Endocrine Society Clinical Practice Guideline as indicative of remission [16]. A more strict day 3 cut-off of 50 nmol/L (1.8 μg/dl) has been reported in paediatric studies [17], and also included in the Endocrine Society Guideline [16]; the literature suggests this cut-off is associated with remission, and a low recurrence rate of approximately 10% at 10 years [14]. The main objective of this study was to assess the outcomes of endoscopic transsphenoidal surgery for Cushing’s disease in a tertiary pituitary centre; remission using two widely accepted criteria [16], recurrence and postoperative complications.

Reported remission rates following ETSS in patients with Cushing’s disease (CD) vary widely, predominantly due to differences in criteria used to define remission [11]. There is no uniform consensus on the criteria used to define ‘remission’, with institutions using a combination of biochemical and clinical criteria; this makes comparing surgical outcome studies challenging. The normal corticotroph cells of the pituitary gland are suppressed due to sustained hypercortisolaemia, therefore following successful removal of the ACTH-secreting adenoma, serum ACTH and cortisol concentrations should fall postoperatively. A morning serum cortisol concentration < 138 nmol/L (5 μg/dl) within 7 days of ETSS is usually indicative of remission, and this biochemical cut-off is quoted in the Endocrine Society Clinical Practice Guideline [16], and many surgical outcome studies [8, 11, 25]. Other studies have applied a more strict serum cortisol cut-off of < 50 nmol/L (1.8 μg/L) at day 3 postoperatively to allow early indication of biochemical remission [10, 11, 26‐28]; the literature suggests this cutoff is associated with remission, and a low recurrence rate of approximately 10% at 10 years [14]. Our practice is to apply this latter approach; if serum cortisol on day 3 is 50–138 nmol/L, serial measurements are taken daily to determine if cortisol will fall further, and assessment for improvement/resolution of clinical signs of hypercortisolaemia made, before repeat endoscopic transsphenoidal surgery is considered. It is important to ensure that serum cortisol has reached a nadir, before further intervention is considered.

In this single-centre single-surgeon study, we report two very different remission rates using these two widely accepted criteria. Our remission rate, including those patients who had an early second ETSS, using standard guidelines, is 92%, on par with other larger studies [7, 8, 11, 25, 29]. When patients with corticotroph macroadenomas were excluded, the remission rate was even higher at 94%. In comparison, when we applied the more strict criteria of day 3 cortisol < 50 nmol/L, the remission rate was considerably lower at 61%. This criteria is in place in our institution so that we can safely identify patients who have early signs of remission to facilitate discharge on day 3 post-operatively; however reporting these rates in isolation lead to a misleadingly low remission rate compared to the more lenient criteria proposed by the Endocrine Society [16].

Evidence has suggested that higher day 3 cortisol concentration is associated with greater risk of recurrence of CD. A recent retrospective cohort analysis of 81 ETSS for CD by Mayberg et al. reported significantly higher recurrence rates in patients with post-operative cortisol nadir between 58 and 149 nmol/L (2.1–5.4 μg/dL) compared with those with cortisol < 55 nmol/L (2 μg/dL) (33% vs 6%, p = 0.01) [30]. Recurrence of CD was low in our series at 3%, and occurred in a patient with a corticotroph macroadenoma, which have been shown to be associated with higher rates of recurrence [31]. On post-operative assessment, serum cortisol fell between the two criteria for remission and if remission was strictly defined as a day 3 cortisol < 50 nmol/L, then this patient had in fact persistent hypercortisolaemia. This case highlights the difficulty when comparing studies reporting ETSS outcomes in CD – the distinction between persistent post-operative hypercortisolism and early recurrence of CD is not always clear-cut, and is dictated by the local protocol.

Whilst our recurrence data are encouraging in comparison to other reports on CD recurrence, which published rates of up to 22% [11], longer term follow-up is necessary before recurrence rates can be accurately defined. The criteria used to define long term recurrence of CD also varies widely in the literature; a large systematic review (n = 6400) by Petersenn et al. (2015) reported decreased recurrence rates when studies used UFC with ONDST vs. UFC only, and UFC with morning serum cortisol vs. UFC only [11]. This highlights the requirement for standardization of remission and recurrence criteria, for consistency in clinical practice and in the literature.

The post-operative surgical complication rate in our series was very low, with no cases of CSF leak, vascular injury or visual compromise. Other published case series have reported incidence rates for CSF leakage and meningitis of 0–7.2% and 0–7.9% [2, 12, 32, 33] respectively. Postoperative meningitis is strongly associated with CSF leakage [34]. Some studies suggest that the endoscopic approach results in higher rates of carotid artery injury compared with the microscopic approach, which could be attributed to the nature of the extended lateral approach [35]. However, in this series of 43 ETSS, we report no cases of surgical related carotid artery injury, similar to other studies reporting 0% serious morbidity or mortality due to carotid artery injury [33, 36]. Finally, postoperative visual disturbance is a major concern, as it can be life changing for patients. Factors linked with visual complications include tumour size, patient age and any pre-existing visual conditions [37‐39]. Visual deterioration after TSS for Cushing’s disease has been reported to occur in some large case series at rates of 1.9% [32] and 0.86% [12]. There were no cases of postoperative visual disturbance in our series.

While the surgical complication rate was low, our endocrine complication rate was higher than that reported in other studies, particularly the rate of DI. Transient DI occurred in 33% of cases, and permanent DI in 23%. These relatively high rates of transient DI may be due to the diagnostic criteria used in our protocol; we defined transient post-operative DI as one episode of hypotonic polyuria in the setting of normal or elevated plasma sodium concentration, requiring at least one dose of desmopressin. In contrast, some studies discount any polyuria which lasts less than 2 days [10], while others require the documentation of hypernatremia for the diagnosis of DI [40]. These more stringent criteria will not capture cases of mild transient DI; therefore it is not surprising that the rates of transient DI reported in a 2018 meta-analysis were lower than that in our study, 11.3% [29]. The rates of permanent DI in our study merits particular attention. TSS for CD has been shown to be associated with a higher risk of post-operative DI [41, 42]. It may be that a more aggressive surgical approach resulted in high remission rates, but at a cost of higher rates of DI. All patients are reviewed post-operatively in the National Pituitary Centre, where there is a low threshold for water deprivation testing and/or 3% saline testing. We did not routinely re-test patients for resolution of DI after their initial water deprivation test at 3 months, and it is possible that some cases subsequently resolved after 3 months [41, 43]. Regardless, the rate reported in this study is significant, and emphasises the importance of counselling the patient about the risk of DI long-term.

Endoscopic transsphenoidal surgery in patients with Cushing’s disease offers excellent remission rates and low morbidity. Remission rates are much higher when standard criteria [morning serum cortisol < 138 nmol/L (5μg/dl) within 7 days postoperatively] are used compared with day 3 cortisol < 50 nmol/l. Higher remission rates were found for patients with microadenomas. Patients should be counselled regarding risk of post-operative endocrine deficiencies, in particular permanent diabetes insipidus. Longer follow-up is required to accurately assess recurrence rates.