Postoperative Hormonal Outcomes in Patients with Large and Giant Non-functioning Pituitary Adenomas
Correspondence Address: Source of Support: None, Conflict of Interest: None DOI: 10.4103/0028-3886.287670
Source of Support: None, Conflict of Interest: None
Keywords: Diabetes insipidus, endoscopic surgery, hormonal axis, pituitary adenoma, risk factorsKey Message: Hormonal dysfunction which is seen in nearly two-thirds of patients with large and giant pituitary adenomas remains the same in around 70% of patients following surgery.
Large pituitary macroadenomas are known to cause visual and pituitary dysfunction. After surgery, 80% to 90% of patients have a good recovery of vision, but the recovery of pituitary function is much less than that of vision.,, The rate of postoperative pituitary function improvement has been reported to vary from 0% to 60%.,,,,,
There are very few studies on pituitary function, before and after surgery in patients with large and giant pituitary adenomas.,,,,, Comparison of recovery of pituitary function or hormonal axis (HA) function following surgery, reported in different published series, is challenging due to differences in postoperative hormone replacement protocols and differences in the tumor size in different studies. We designed this study to analyse the risk factors and outcomes for pituitary dysfunction preoperatively and at three months after surgery in patients with large (Hardy's grade C) and giant nonfunctioning pituitary adenomas (NFPA).
All patients with large (>3 cms) (Hardy's grade C) and giant NFPA operated between 2006 and 2017, in whom pre-operative and three months post-operative HA functions were available, were identified. HA function data before and three months after surgery were collected.
Tumour volume calculation
Total tumor volume (TTV) was calculated on preoperative and postoperative MR images. The maximum diameter of the tumor in 3 axes perpendicular to each (a, b, c) was measured and the TTV was calculated using the formula “a × b × c/2”.
Perioperative hormonal replacement protocol
All patients underwent preoperative hormonal tests (8 am serum cortisol, ACTH, TSH, T3, FT4, prolactin, FSH, LH, testosterone, GH, IGF-1). Patients with hypocortisolemia, hypothyroidism or hypogonadism were started on replacements preoperatively. Intraoperatively, all patients received one dose of injection hydrocortisone 100 mg followed by 25 mg every sixth hourly. This was tapered and overlapped with oral prednisolone over the next 2-3 days. Subsequently all patients were discharged on oral prednisolone 5-0-2.5 mg twice daily. Over the next three months, prednisolone was tapered to 5 mg on alternate days. At a follow-up done three months later, prednisolone was withdrawn under supervision for 48 hours to assess the recovery of cortisol axis. Those with inappropriately low cortisol levels (<5 ug%) were asked to continue prednisolone long term.
The surgery was performed entirely with endoscopes and endoscopic instruments. A uni-nostril trans septal approach was used to perform the surgery. A posterior septal incision was made to elevate septal mucosal flaps on either side of the nasal septum. The posterior nasal septum was excised partially till the rostrum of the sphenoid sinus was reached. A thin blade tapered Killian's nasal speculum was used to keep the septal flaps apart. Using a long telescope (30 cm) held on a table mounted endoscope holder (Karl Storz, Germany) bi-manual surgery was possible and this facilitated excision of the tumor. The tumor was excised using ring curettes until the diaphragm descended into the sella. Angled suctions, ring curettes and angled endoscopes were used to ensure that a radical excision was achieved. We attempted to preserve the normal gland in all cases.
The sella and sphenoid sinus were packed with abdominal fat and oxidized cellulose and gelatine sponge. In case of a breach in the diaphragm and intra-operative CSF leak, the breach was repaired with the same materials. The nose was not routinely packed.
HA data collection
Hormonal evaluation was done before and three months after surgery. HA dysfunction was graded as 0, 1, 2 or 3 depending on the number of axes involved, this included the thyrotroph axis, the corticotroph axis and the gonadotroph axis.
IHC was performed on 121 biopsy samples and correlated with the preoperative and postoperative hormonal axis function.
The following risk factors were studied for their correlation with the HA functions: age, tumor volume, diabetes mellitus, hypertension, duration of symptoms and extent of resection (for post-operative HA function only). For continuous variables, descriptive statistics such as mean, standard deviation (SD) and for abnormally distributed interval data and ordinal data, median (interquartile range [IQR]) were calculated. Number of patients and percentage was presented as categorical data. Chi-square and Fisher's Exact test was used to find association between categorical variables. Pearson correlation was used to find correlation between continuous variables. Logistic regression was used to determine the association of a risk factor with the outcome. All tests were two-sided at α = 0.05 level of significance.
Receiver operator characteristic (ROC) curve was plotted against true positive (sensitivity) versus false negative (1- specificity) at each tumor size threshold for pan hypopituitarism (dysfunction of all three hormonal axes). Delong et al. method was used for plotting area under the curve (AUC) with a binomial exact confidence interval for AUC. Each point on the ROC curve represents a sensitivity/specificity pair corresponding to a particular threshold. Therefore, closer the ROC curve to the upper left higher the overall accuracy of the test.
All analyses were done using Statistical Package for Social Services (SPSS) software Version 21.0 (Armonk, NY: IBM Corp).
Demographics and symptoms
There were 139 patients, M:F ratio was 2:1 and mean age was 44 years (SD, 13.0). Most patients (100, 72%) were <50 years of age. The average tumor size and median volume were 3.6 cms and 15 cc (range 3-96 cc) respectively [Table 1]. Majority of patients presented with visual dysfunction (84%). The other symptoms included headache, seizure and easy fatigability. Only 4.7% of patients presented with pituitary apoplexy. Nearly half of the patients (49%) had symptoms for <6 months, while 12% of patients had symptoms for >2 years.
Most of the patients (89, 64%) did not have any comorbidities. Diabetes mellitus was present in 11 (8%) patients and hypertension in 29 (21%) patients while 10 (7%) patients had both.
Preoperative HA function
Out of 139 patients, 45 (32.3%) had no HA dysfunction, 34 (24.4%), 36 (25.8%) and 24 (17.2%) had one, two and three axes involved, respectively. None of the patients had diabetes insipidus (DI). Out of the three axes, thyrotroph axis (n = 88, 63.3%) was most commonly affected followed by corticotroph axis (n = 55, 39.5%) and gonadotroph axis (n = 35, 25%) [Table 2].
Risk factors for pre-operative pituitary dysfunction
Symptom duration (>6 months) and preoperative tumor volume were significantly associated with HA dysfunction (P = 0.01), in univariate analysis. However, in multivariate analysis, tumor volume was the only statistically significant factor (P = 0.01), that was associated with HA dysfunction. Higher the tumor volume, greater the risk for hormonal dysfunction [Table 3]. ROC curve analysis showed significant correlation (P< 0.000) with area under curve (AUC) value being 0.701 with standard error of 0.05 at 95% confidence interval [Figure 1]. At the threshold volume of 15.0 cc, the sensitivity and specificity for pan hypopituitarism (>2 HA dysfunction) were 0.79 and 0.57 respectively. The positive predictive value (PPV) and negative predictive value (NPV) were 0.27 and 0.93 respectively.
Postoperative HA function
Post operatively, pituitary function remained the same, as before surgery, in 100 patients (72%) it improved in 7 (5%) patients and worsened in 32 (23%). Post-operatively, normal pituitary function was seen in only 17.3%, a drop of 15% from the preoperative level of 32.3% [Table 2]. The percentage of patients with 1 and 2 HA dysfunction increased by 2.8% and 7.1% respectively. Percentage of patients with panhypopituitarism increased from 17.3% to 20.1% an increase of 2.8%. Of those with panhypopituitarism at follow-up (n = 28), 24 (85.7%) had panhypopituitarism before surgery, 3 (10.7%) had two axes involved and 1 (3.57%) had one axis involved. Rate of worsening of HA was higher in the corticotroph axis than thyrotroph axis. Corticotroph axis dysfunction which was seen in 39.5% of patients before surgery, was noted in 60.4% of patients after surgery. Thyrotroph axis dysfunction which was seen in 63.3% of patients before surgery, was noted in 69.7% of patients after surgery. Gonadotroph axis dysfunction was seen in slightly fewer patients at the follow up visit (23.7% at follow up versus 25.2% preoperatively). Six patients (4.3%) had permanent DI.
Risk factors for post-operative HA function
A larger percentage of patients <50 years of age maintained their HA function (77%) while 6% of patients showed improvement compared to patients ≥50 years of age (39). In the latter group, 23 (59%) retained their HA function and only one patient (2.5%) improved. However, this difference was not statistically significant [Table 4]. Worsening of pituitary function was more common in patients with underlying comorbidities like DM and HTN (56.3%) as compared to those without these comorbidities (43.8%). Again, the difference was not statistically significant.
The proportion of patients with HA worsening after GTR or NTR of tumors was 21.8% while 27.8% of patients undergoing STR/partial excision/biopsy had HA worsening. Extent of resection did not have any statistically significant correlation with the hormonal outcome [Table 4].
IHC revealed gonadotroph (FSH, LH) positive tumors in 71 (51%) patients, null cell adenomas in 29 (20.8%) patients, silent corticotroph adenomas in 11 (8%) patients, mixed IHC positive adenomas in 9 (6.7%) patients, thyrotroph adenoma in 1 (0.7%) patient. IHC status was not known in 18 (12.9%) patients. All the tumors were non-secretory. The postoperative HA outcome was similar in all the adenoma sub-types [Table 5].
The most common postoperative complication was hyponatremia (15%). Twenty-one patients developed DI. In 15 patients it was transient and permanent in 6. Postoperative CSF leak occurred in two patients of whom one required re-operation to repair the leak.
Preoperative pituitary function
In literature, patients with large and giant adenomas have been reported to have greater incidence of pituitary dysfunction when compared to those with smaller tumors.,, The incidence of pituitary dysfunction in small tumors (<3 cms) range from 40% to 70%.,,,, Whereas 60% to 90.3% of patients with large and giant NFPA were noted to have partial or complete hypopituitarism.,,,, In our study 94/139 (67.6%) patients had either partial or complete hypopituitarism [Table 6]. Several authors have reported that larger tumor volumes are associated with greater degree of pituitary dysfunction. Zhang et al. reported the threshold tumour volume to predict hypopituitarism was 3.1 cc (sensitivity and specificity of 89.0 and 67.3%, respectively). Patients with tumours >3.1 cc were more likely to have pituitary dysfunction (one or more axis). In our study, we found 15 cc as threshold volume for preoperative panhypopituitarism with P value <0.0001 (sensitivity and specificity of 79.2 and 57.4).
Dysfunction rates for corticotrophic, thyrotrophic, gonadotrophic and somatotrophic axes, varies in different series.,,,,, Chen et al. and Rother et al. showed higher dysfunction in somatotroph axis whereas Iglesias et al., Fallah et al. and Musluman et al. showed higher dysfunction in gonadotroph axis. Our study showed higher dysfunction in thyrotroph axis and this is similar to that reported by Jahangiri et al. and Laws Jr et al. [Table 6].
Nishioka et al. in their series of giant NFPA (n = 128, mean size 4.8 cms) had 5 patients (3.9%) with preoperative DI. Preoperative DI was not seen in the other series,, as well as in our study. Hence preoperatively, posterior pituitary function is seldom affected.
Postoperative pituitary function
Fallah et al. noticed 5-15% of improvement in different HA with maximum improvement being noted in the corticotroph axis (15%). Hormonal axes remained unchanged in 75-80% of patients, while it worsened in 16.4% of patients. Our findings are similar to those of Fallah et al., Nishioka et al. and Iglesias et al. who reported that between 70-90% of patients will maintain their HA function following surgery [Table 7].
Improvement of pituitary function
The reported postoperative hormonal recovery in large and giant NFPA range from 0% to 44%.,,, Most authors have reported improvement in HA function in 0% to 10% of their patients.,, The outlier is the series of Musluman et al. who reported improvement in 44% of their patients. Preoperative hormonal dysfunction in their series (60%) was similar to that seen in our series (67.6%) and therefore, it is difficult to explain the better hormonal outcomes in their series. One difference is that surgery was performed within six months of onset of visual symptoms in 71.8% of their patients as compared to 49% of our patients. But duration of symptoms did not correlate with HA function outcomes in our series.
Risk factors for the outcomes of pituitary function
Nomikos et al., in his large series of 720 patients (micro and macroadenomas) with NFPA, found no correlation between age and hormonal outcomes. In our series also, there was no significance found between age and the hormonal outcome.
Zhang et al. reported that postoperative hormonal recovery was poor in giant adenomas, although the exact percentages and the factors influencing the recovery were not reported.
Nomikos et al. correlated tumor size with pituitary recovery and found higher recovery with smaller tumors (P < 0.001). The mean size of the tumor associated with pituitary function recovery or improvement was 26 ± 6 mm. This size was significantly less when compared with that of tumors with no improvement in pituitary function (33+/- 9 mm, P < 0.001). It was difficult to compare our results with these studies due to differences in the size of tumors included in the analysis. Nomikos et al. included patients with tumors which ranged in size from microadenomas to macroadenomas whereas we included patients with only large (>3 cms) sized tumors. Therefore the 2.6 cm cut off for the tumor size associated with postoperative hormonal improvement is less than the smallest tumor (3 cms) included in our study.
In our series, we did not find any correlation between the tumor size and recovery of pituitary function.
Comorbidities (DM and HTN)
None of the reported studies has evaluated the significance of comorbidities like DM and HTN on hormonal outcomes. We did not find any significant correlation of these risk factors with the hormonal outcomes.
Extent of resection (EOR)
There is no literature available on the extent of resection and outcomes of pituitary function in patients with large and giant NFPA. We speculated that a more radical excision of the tumor would be associated with a higher risk of damage to the compressed pituitary gland and therefore, patients undergoing GTR or NTR might have a higher rate of pituitary dysfunction in the post-operative period. However, a slightly higher proportion of patients undergoing STR or biopsy had worsening of HA dysfunction (27.5% vs. 21.8%). We did not find any correlation between EOR and postoperative HA function.
Role of Immunohistochemistry (IHC)
In 2017, WHO sub classified NFPAs into 8 types based on IHC expression of adenohypophysis and transcription factors with the recent proposed name as non-functioning pituitary neuroendocrine tumors (NF-PitNETs). The main objective behind this classification was 1) For better understanding the biology of these tumors; 2) Identification of factors involved in invasiveness and recurrence; 3) Identification of clinically aggressive subtypes; 4) Identification of predictive markers which will help in developing new pharmacological therapeutic strategies, and better selection of patients suitable for radiation therapy. Insufficient standardization of methodology, lack of specific antibodies and inappropriate interpretation of IHC staining makes it difficult to phenotype these tumors. Molecular markers are not added in the current classification because of unclear genetic pathways.
There are very few studies which tried to identify the significance of IHC subtypes and proliferative index in the invasiveness of the tumor, recurrence rate and outcomes.,, Most of the giant adenomas are invasive and surgically challenging, but most of them are histologically benign, slow growing and gonadototrophic in subtype., Nishioka et al. and few other authors, compared the outcomes and found silent corticotrophs to be more clinically aggressive than the gonadotrophs.
In our study, we found no correlation between the IHC subtype and the preoperative hormonal dysfunction; this was similar to what was observed by Nishioka et al. and Iglesias et al.
In large (Hardy's C) and giant NFPA, preoperatively anterior pituitary function is affected in 67.6% of patients and posterior pituitary function is usually spared. Preoperative tumor volume >15 cc is associated with pan-hypopituitarism. Postoperatively, 72% of the patients retain their preoperative pituitary function with improvement seen in only 5% of patients. No significant risk factors were found for the postoperative hormonal outcomes.[28-32]
Financial support and sponsorship
Conflicts of interest
There are no conflicts of interest.
[Table 1], [Table 2], [Table 3], [Table 4], [Table 5], [Table 6], [Table 7]