Patients
Among the 1,122 pituitary adenomas, 136 (12.1%) demonstrated biochemical evidence of acromegaly and 124 stained positive for GH on histology. Sixty-two men (45.6%) and 74 women (54.4%) were diagnosed at mean age of 48.1 ± 13.8 years. A majority of patients reported onset of symptoms at mean age 40.5 ± 13.8 years, with average diagnostic delay of 7.6 years. Available mortality data revealed 94.1% of patients were alive at the time of data collection based on our institution’s electronic health records.
Table 1 details the demographics and presenting symptoms. Musculoskeletal complaints (acral or facial enlargement, dental problems, osteoarthritis, carpal tunnel syndrome) were reported by 89.6% of patients, compressive symptoms (headache, visual changes or cranial nerve deficits) by 65.6% of patients, and hypogonadism (amenorrhea, sexual dysfunction, gynecomastia, galactorrhea) by 45.2% of patients. Miscellaneous concerns including hyperhidrosis, acne, hirsutism and dizziness were reported by 58.3% of patients, while 40.6% complained of psychological symptoms including mood changes and depression. Only a minority of patients (12.1%) were incidentally discovered to have a sellar mass on imaging as the presenting feature.
Patients presented with comorbidities including hypertension (56.7%), obstructive sleep apnea (47.5%) and diabetes mellitus (44.4%). Cardiac dysfunction including coronary artery disease, congestive heart failure or cardiomyopathy was observed in 16.8% of patients, while 12.9% had a history of, or were subsequently diagnosed with, malignancy (Table 1). Multiple endocrine neoplasia type-1 was confirmed with genetic testing in one patient.
Table 2 details the biochemical findings and tumor characteristics on presentation. Median (minimum to maximum range) IGF-1 level at diagnosis was 769.0 ng/mL (235 to 1845 ng/mL), with median baseline GH level of 8.2 ng/mL (0.17 to 808.12 ng/mL), and median nadir 2-hour post-OGTT GH level of 4.4 ng/mL (0.2 to 330.0 ng/mL). Most patients were discovered to have a macroadenoma (82.2%), with median tumor dimensions of 16.0 (antero-posterior) x 16.0 (transverse) x 14.0 (craniocaudal) mm. CS invasion (44.7%) and OC compression (32.6%) were noted on MRI. GH staining was positive in 124 adenomas (91.2%). Out of 65 adenomas with reported staining pattern, 36 demonstrated diffuse staining for GH (55.4%), while 29 demonstrated rare GH staining pattern (44.6%). Pituitary hormone co-staining was observed in 89 adenomas (65.4%), including 57 adenomas staining for both TSH and prolactin (PRL), 12 staining for ACTH, and 4 staining for FSH. Cytokeratin CAM5.2 staining performed in 23 patients revealed 17 DG and 6 SG adenomas.
Treatment and Outcomes
Eighty-four patients (62.7%) received pre-surgical medical treatment as follows: octreotide (80.9%) or lanreotide (7.1%) alone, followed by combination therapy with octreotide and cabergoline (4.8%), while 2 patients received PEG (Table 3). All 136 patients underwent pituitary tumor resection; four patients had initial craniotomy (2.9%), while 132 patients underwent transsphenoidal resection (97.1%). Complete visible tumor resection during initial surgery was achieved in 87 patients (64.0%). Subsequently, 19 patients required repeat surgery (of which three were open craniotomy), and 27 patients underwent irradiation.
Medical therapy post-surgery was initiated in 65 patients (47.8%), with 32 patients receiving SSA alone (49.2%), 3 patients receiving DA alone (4.6%), while 11 received combination therapy with SSA, DA and PEG (16.9%), and 1 patient received SSA, DA, PEG and pasireotide in step-wise escalation of therapy (Table 3). Complete visible tumor resection during initial surgery was achieved in 87 patients (64.0%). In this cohort, complete response to surgery alone was observed in 61 patients (70.1%), while 31 patients out of 65 (47.7%) who received subsequent post-surgical medications and/or radiation therapy achieved complete response. At most recent follow-up, 92 out of the total cohort of 136 patients (67.6%) achieved eventual complete response by documented normalization of IGF-1 levels. Overall, a significant majority of patients reported improvement in symptoms (91.5%) and had objective improvement in comorbidities (69.0%) after acromegaly treatment. The development of postsurgical hypopituitarism was noted in 22 out of 133 patients (16.5%). All patients who developed hypopituitarism had a history of irradiation, and 16 out of 22 patients (72.7%) with hypopituitarism had undergone repeat surgery.
Univariable Analyses with Odds Ratios
Table 4 details the outcomes of univariable analyses based on completeness of initial surgical tumor resection. Patients with complete initial resection of visible tumor had lower GH levels at diagnosis (P = 0.001) and smaller tumors in all dimensions (all P < 0.001). For each 1 ng/mL unit increase in GH, the odds of complete tumor resection decreased by 2% (OR (95% CI): 0.98 (0.97, 1.00), P = 0.001). Those with complete initial resection were also more likely to have GH staining (P = 0.022) and pituitary hormone co-staining (P = 0.026). They were less likely to have CS invasion and OC compression (both P < 0.001) or develop hypopituitarism (P < 0.001). The odds of a patient with hypopituitarism having had complete initial resection were 92% less than a patient without hypopituitarism (OR (95% CI): 0.08 (0.03, 0.27), P < 0.001).
Outcomes of univariable analyses based on response to surgery in those with complete initial resection were performed. Complete response to surgery was defined as normalization of IGF-1 levels without evidence of residual tumor at the time of last follow-up. Patients who responded to surgery had lower GH (P = 0.027) and lower IGF-1 levels (P = 0.001) at diagnosis, along with lower GH level (P = 0.031) and 2-hour nadir post-OGTT GH level (P = 0.014) at most recent follow-up. Lastly, patients who responded to surgery were less likely to have CS invasion (P = 0.015). Outcomes of univariable analyses based on response to medication in those with complete initial resection were performed. Response to medication was defined as normalization of IGF-1 levels at last follow-up. Patients who responded to medication had lower levels of GH (P = 0.003) and IGF-1 (P = 0.046) at the time of their most recent follow-up.
Outcomes of univariable analyses of those who underwent repeat surgery were performed. Patients with repeat surgery were younger (P = 0.026), had higher GH levels at diagnosis (P = 0.011), larger tumors (anteroposterior P < 0.001, transverse P = 0.002, and craniocaudal P = 0.002), with CS invasion (P = 0.044) and OC compression (P = 0.011). Patients who underwent repeat surgery were less likely to have pituitary hormone co-staining with TSH (P = 0.047). Repeat surgery increased the risk of hypopituitarism (P < 0.001), with hypopituitarism present in 16 out of 19 patients (84.2%).
Outcomes of univariable analyses based on response to medication in those who had incomplete initial tumor resection were examined. Patients who responded to subsequent medical therapy were more likely to be female (P = 0.008), older at onset of symptoms and age at diagnosis (both P < 0.001), and less likely to report compressive symptoms on presentation (P = 0.022). These patients also had lower GH level (P < 0.001) at most recent follow-up. Outcomes of univariable analyses based on improvement in symptoms and comorbidities were performed. Patients who reported improvement in symptoms were more likely to be older at onset of symptoms (P = 0.036) and age at diagnosis (P = 0.023), and less likely to have hypopituitarism (P = 0.046). Patients who had objective improvement in comorbidities were more likely to be female and older at diagnosis (both P = 0.0.023). They had smaller tumor transverse dimension on presentation (P = 0.008) and less likely to have CS invasion (P = 0.008) and OC compression (P = 0.029). These patients also were more likely to have tumors stain positive for GH (P = 0.015) and co-stain with TSH (P = 0.036). Patients with improvement in comorbidities were less likely to have hypopituitarism (P < 0.001).
Subgroup analyses in Table 5 reveals that patients with SG adenomas were more likely to require repeat surgery than those with DG adenomas (P = 0.040).
Multivariable Analyses
Multivariable models for incomplete surgical resection and failure to respond to surgery among those with complete resection were analyzed. Predictors for failing to achieve complete initial tumor resection are CS invasion (P = 0.048) and OC compression (P = 0.023). Patients with CS invasion have 2.4 times higher odds of failure to achieve complete resection compared to patients without CS invasion, after adjustment for other terms in this model. Additionally, subsequent development of hypopituitarism is statistically correlated (P < 0.001) with incomplete initial tumor resection. Predictors of failure to respond to surgery after complete initial visible tumor resection are higher IGF-1 level at diagnosis (P = 0.024) and CS invasion (P = 0.028). Significantly, for each 10 ng/mL unit increase in IGF-1 at diagnosis, the odds of failure to respond to surgery increased by 2%, after adjusting for other factors. The development of hypopituitarism (P = 0.047) is correlated with failure to respond to surgery.