A Prospective, Multicenter, Observational Study of Surgical vs Nonsurgical Management for Pituitary Apoplexy

Abstract

Context

Pituitary apoplexy (PA) has been traditionally considered a neurosurgical emergency, yet retrospective single-institution studies suggest similar outcomes among patients managed medically.

Objective

We established a multicenter, international prospective registry to compare presentation and outcomes in PA patients treated with surgery or medical management alone.

Methods

A centralized database captured demographics, comorbidities, clinical presentation, visual findings, hormonal status, and imaging features at admission. Treatment was determined independently by each site. Key outcomes included visual, oculomotor, and hormonal recovery, complications, and hospital length of stay. Outcomes were also compared based on time from symptom onset to surgery, and from admission or transfer to the treating center. Statistical testing compared treatment groups based on 2-sided hypotheses and P less than .05.

Results

A total of 100 consecutive PA patients from 12 hospitals were enrolled, and 97 (67 surgical and 30 medical) were evaluable. Demographics, clinical features, presenting symptoms, hormonal deficits, and imaging findings were similar between groups. Severe temporal visual field deficit was more common in surgical patients. At 3 and 6 months, hormonal, visual, and oculomotor outcomes were similar. Stratifying based on severity of visual fields demonstrated no difference in any outcome at 3 months. Timing of surgery did not affect outcomes.

Conclusion

We found that medical and surgical management of PA yield similar 3-month outcomes. Although patients undergoing surgery had more severe visual field deficits, we could not clearly demonstrate that surgery led to better outcomes. Even without surgery, apoplectic tumor volumes regress substantially within 2 to 3 months, indicating that surgery is not always needed to reduce mass effect.

Pituitary apoplexy (PA) is a relatively rare but well-recognized syndrome caused by rapid hemorrhage and/or infarction of a tumor in the pituitary fossa (1). PA presents with a constellation of symptoms including headache, nausea, vomiting, visual changes, oculomotor palsies, and hormonal deficiencies. PA, first described in 1898 (2) and characterized as a pathological entity in 1950 (3), is traditionally considered a neurosurgical emergency requiring surgical decompression and corticosteroid administration (4-7). Although many studies report excellent return of visual function and resolution of oculomotor palsies after surgery (1), others (8-10) suggest that medical management alone without surgery is equally effective and yields similar outcomes at 3 months.

However, these reports are single-center, retrospective outcomes reviews (11-14), are incomplete or focus narrowly on patient outcomes and/or presenting features (5, 8, 9, 15-20), and do not necessarily consider surgical expertise, which is known to influence outcomes (21).

To address the weaknesses of existing studies, we established the Pituitary Apoplexy Surgical Timing and Outcomes Registry (PASTOR), a multicenter prospective international registry designed to collect uniform outcome results on patients managed with or without surgery at centers with expertise in management of pituitary disease. We sought to: (1) identify differences in patient populations managed medically or surgically; (2) evaluate short- and long-term clinical and quality-of-life (QoL) outcomes; and (3) measure whether time from onset of symptoms to surgery influenced outcomes.

Materials and Methods

Study Approval and Design

This prospective registry was approved by the Cedars-Sinai Institutional Review Board (IRB No. 49637), which served as the central IRB of record. A data-sharing agreement was executed between Cedars-Sinai and each site. Data were anonymized prior to collection using REDCap, a web-based platform that supports research data capture (22, 23).

Enrollment was determined by the investigators at each site without central oversight. All patients older than 18 years with a diagnosed PA based on established clinical and radiographic features were eligible. Patients with hemorrhage or infarction identified on imaging studies but with no clinical symptoms of PA were not included. Treatment decisions were made independently at each site based on investigators’ best judgment and treatment preferences. There was no central review of radiology, pathology, or other data. Consent was obtained by each site, and enrollment in the study did not affect management decisions.

Assessments

Visual acuity and visual fields

Semiquantitative scales measured visual acuity and visual fields based on bedside examination. The ability to read a smartphone was considered grossly normal acuity, and counting fingers but not reading a smartphone, followed by seeing a hand wave or only light perception, were graded as progressive levels of diminished acuity. Only the best score was provided, so that if a patient could read a smartphone, it was understood they could also count fingers. Visual fields were scored as normal, mild temporal defect, severe temporal defect, or blind. If there was a discrepancy in severity between eyes, the more severe score was assigned.

Oculomotor palsy

Oculomotor palsies were scored as none, partial, or complete (ie, no oculomotor nerve functioning, resulting in a frozen eye). Left and right eye data were collected separately but grouped for analysis.

Radiographic measures

Tumor diameter was defined as the largest cross-sectional distance in any plane on pituitary MRI. Tumor volume was estimated using the (X × Y × Z)/2 method. Extent of pituitary gland and optic nerve compression was rated by the site investigator as none, moderate, or severe.

Laboratory measures

Pathology and pituitary hormonal function (intact, deficient, or uncertain) were reported by each site without central review or standardized criteria.

Pituitary Apoplexy Score

Pituitary Apoplexy Score (PAS), used by some to grade clinical severity of the apoplectic event, was calculated for each patient according to published criteria (Table 1) (24). Briefly, visual acuity, visual field deficits, oculomotor palsy, and level of consciousness are individually scored then combined to describe PA severity on a 0 to 10 scale, with 10 indicating worse function. While PAS was developed based on retrospective data and is not validated to predict outcomes, it has been proposed as a means to identify patients more likely to require surgery.

Data Entry

Data entry was managed by each site and reviewed at Cedars-Sinai for incomplete data. Patient demographics, symptoms at presentation, comorbidities, surgical vs medical management, visual, oculomotor, and hormonal function, and radiographic findings at presentation were recorded, as was pathology for surgical patients.

Hospital length of stay (LOS), LOS in the intensive care unit (ICU), hospital-related complications of treatment, including readmission rates, as well as the modified 5-point frailty index (mFI-5) (25) were recorded for all patients. Time from symptom onset to surgery and time from admission to surgery were recorded for surgical patients.

Hormonal, visual, and oculomotor function, as well as patient responses to the Headache Impact Test (HIT-6) (26) and the Short Form Health Survey (SF-36) QoL survey (27) were recorded at 3 and 6 months.

Statistical Plan

Continuous variables were summarized using medians and interquartile ranges (IQRs), and categorical variables using frequencies and proportions. Wilcoxon rank sum test was used to test for continuous variables between treatment groups and Pearson chi-square or Fisher exact tests for categorical variables. Hypotheses were 2-sided with a P value less than .05 considered statistically significant. Statistical analyses were performed using R (version 3.6.3, R Core Team). As there is a consensus among neurosurgeons that apoplexy requires surgery, we assumed a 2:1 ratio of surgical vs medical management in the overall cohort. Power analysis indicated that collection data on 100 patients, of which 66% were surgical and 33% were medical, would provide an 80% chance of differentiating improved visual or hormonal outcomes at 3 months with statistical significance.

Results

One hundred consecutive patients were enrolled in 12 hospitals between November 2018 and February 2022, including 8 hospitals in the United States, 1 in Canada, 1 in France, 1 in Japan, and 1 in Korea (Fig. 1). Three patients were excluded for insufficient data. Of the remaining 97 evaluable patients, 67 (69%) were treated with surgery and 30 (31%) were treated without surgery. Follow-up data were complete at 3 months, but less so at 6 months. Therefore, we present here our findings from the 3-month data and draw mostly qualitative conclusions for the 6-month data.

Patient Characteristics at Initial Presentation

The surgical group comprised 23 (34%) women and 44 (66%) men, while the medical group comprised 6 (20%) women and 24 (80%) men (Table 2). Median age was 53 years (IQR 40-62 years) in the surgery group and 50 years (IQR 35-72 years) in the medical group. No significant age-related differences were noted. In both groups, 57% arrived at the treating center by direct admission and 43% were transferred after initial evaluation or admission at another hospital. A greater percentage of surgical vs medical patients were working prior to onset of PA (55% vs 20%, respectively, P = .02). mFI-5 frailty score did not correlate with selection of treatment approach.

Symptoms at onset were similar (see Table 2). Headache was the most common symptom (94% surgical and 87% medical), followed by nausea, vomiting, and visual deterioration. There were more cases of vision loss and oculomotor palsy in the surgical group, although this difference was not statistically significant. Body mass index and comorbidities including heart disease, diabetes, and renal insufficiency did not predict treatment decision. Similarly, use of anticoagulants and dopamine agonists did not differ between groups.

Although visual acuity was worse in the surgical group, with 5 (7.5%) reporting profound visual loss vs none reported in the medical group, these differences were not statistically significant. By contrast, visual fields were significantly different between the groups, with surgical patients demonstrating more severe temporal deficits compared to those managed medically (33% vs 10%; P = .017). There were no significant differences in oculomotor palsies between the groups.

Cortisol deficiency was the most common hormonal deficit in both groups, followed by loss of gonadotroph and thyrotropin. No statistical differences between these groups were identified. More patients in the medical group had prolactinomas, but the difference was not statistically significant.

On MRI (see Table 2), there were no significant differences between the groups for maximal tumor diameter (2.6 [IQR 2.1-3.2] vs 2.4 [IQR 1.9-2.8] cm, respectively; P = .12) or tumor volume (5 [IQR 2-10] vs 4 [IQR 2-6] cc, respectively; P = .2). In the surgical group, there were more patients with evidence of optic nerve compression of any degree (moderate or severe, P < .001), and more patients with severe nerve compression (P < .001). The absence of any degree of pituitary gland compression (minimal/none) was more common in the medical group (39% vs 6.5%; P < .001), but there was no statistically significant difference between the groups when comparing severe or moderate pituitary gland compression.

Of note, patients with a PAS of 3 or greater were more likely to undergo surgery (P = .03), and a trend was seen in those with PAS of 4 or greater (P = .053). Thus, at initial presentation, only visual field defects and optic nerve compression on MRI were significantly different between surgical and medical groups, and the higher PAS in the surgical group reflects the more severe visual field defects in these patients.

Perioperative Outcomes

Hospital LOS did not differ between groups; mean LOS was 6.0 ± 4.5 days in the surgery group and 7.0 ± 6.9 days in the medical group (P > .9). However, the surgery group averaged 1 additional day in the ICU (1.5 ± 2.3 vs .5 ± .9 ICU days, respectively).

Of the 67 surgical patients, 65 underwent endoscopic transsphenoidal surgery and 2 had craniotomies. Surgical complications included cerebrospinal fluid leak (n = 3), arginine vasopressin deficiency (n = 3), delayed epistaxis (n = 3), sinusitis (n = 4), and hyponatremia (n = 5). In the medical group, there were 2 cases of arginine vasopressin deficiency and 1 case of hyponatremia.

Within 30 days of discharge, 9 surgical patients (6.1%) were readmitted to the hospital due to cerebrospinal fluid leak (n = 2), hyponatremia (n = 1), epistaxis (n = 1), and conditions unrelated to the PA (n = 5). In the medical group, there was one planned readmission for cardiac testing unrelated to the PA. Nearly all (58/67, 93.5%) surgical patients and all medical patients were discharged to home. One surgical patient was transferred to a skilled nursing facility, and 3 to acute rehabilitation.

Surgical pathology specimens revealed a pituitary adenoma in 63 (94.0%) cases, Rathke cleft cyst in 3 (4.5%), and “other pathology” in 1 (1.5%). “No viable tumor” was reported in 13 cases, necrotic tissue in 34, and hemorrhage in 31. On immunohistochemistry performed in 59 (88%) cases, there were 27 gonadotroph, 4 somatotroph, 9 corticotroph, 2 thyrotroph, and 7 lactotroph adenomas; 10 were graded as null cell, defined as tumors with no positive immunostaining for any pituitary hormone.

Medication Use During Initial Admission

During initial admission, 51 (76%) surgical and 19 (63%) medical patients received corticosteroids, while 33 (49%) surgical and 6 (20%) medical received thyroid replacement (Table 3). The surgical group was more likely to be treated with antibiotics, as expected for surgical prophylaxis. Desmopressin, fluid restriction, and hypertonic saline use were also more commonly used in the surgical group, but these differences were not significant.

Hormonal, Visual, and Oculomotor Function Recovery

No significant differences in hormonal function were observed between the surgical and medical groups at 3 months of follow-up among those with documented deficiencies at presentation (Table 4). Neither surgical nor medical management substantially reversed hormonal deficits. For example, 63% (n = 19) of medical and 66% (n = 44) of surgical patients were cortisol deficient on admission, and only 11.6% (n = 5) and 10.5% (n = 2), respectively, recovered function at 3 months. No additional recovery of any hormone axis was reported at 6 months, although firm conclusions cannot be drawn due to missing data.

Data on visual and oculomotor function were available for 59 of 67 (88%) surgical and 21 of 30 (70%) medical patients at 3 months (Table 5). Oculomotor palsies largely resolved in both groups: 5 (8.5%) patients in the surgical group and 1 (4.8%) in the medical group had partial palsy at 3 months (P > .9), and only 1 patient in the surgical group had complete palsy in at least 1 eye. At 6 months, no oculomotor palsies were reported among the 17 patients with data in the medical group and partial palsy was reported in 4 of the 54 (7.4%) patients with data in the surgical group.

Visual fields were full in 78% of surgical and 71% of medical patients at 3 months (P = .7). At 6 months, visual fields were also largely unchanged from the 3-month assessment, and 1 patient in the surgical group still had a severe temporal field cut at 6 months.

Because severe visual field defects were statistically far more common in the surgical group (33% vs 10%; P = .017), a stratified analysis was performed to determine whether surgery benefited those with severe visual field defects. A total of 67% of patients in the medical group with severe defects at presentation had full visual fields at 3 months compared to 77% in the surgical group (P > .9). No statistically significant difference was seen in any other outcome measure. Similarly, for patients with visual defects graded as “not severe,” there was no statistical difference in visual field outcome (78% vs 72%; P = .6) or any other outcome measure.

Apoplectic Tumor Volume Regression in Medical Patients

MRI data at 2 to 3 months after PA onset were available for 29 of the 30 patients treated without surgery. Four patients were excluded because they underwent elective transsphenoidal surgery after the apoplectic event resolved, but before 3 months after PA onset. An additional patient was excluded because he was treated with cabergoline for prolactinoma, and therefore tumor regression was expected. Of the remaining 24 patients, apoplectic tumor volume regression was noted in all but one (Fig. 2). There was a median change in volume of 3.6(±2.3) cm³ at PA onset to 1.4 (±1.6) cm³ at 2 to 3 months after onset (P = .0002). Similar volumes were not calculated for surgical patients as total or near total removal was reported.

Quality of Life

SF-36 and HIT-6 scores were similar between the surgical and medical groups at 3 months, although QoL was rated slightly lower in the medical group (Table 6). At 3 months, return to work status was similar (81% surgical vs 100% medical; P = .07), although more people in the surgical group were working prior to PA onset. A substantial number of surveys were missing at 6 months. However, based on available data, it did not seem that scores markedly changed from 3 to 6 months.

Outcomes With “Early” and “Late” Surgical Intervention

Some retrospective studies suggest that surgery within 48 to 72 hours of symptom onset improves visual outcomes (19, 29, 30). We therefore evaluated hormonal, visual, and oculomotor function as well as QoL metrics at 3 months in the surgery cohort based both on time from symptom onset (Table 7) and on time from admission or transfer to the treating center (Table 8).

Early vs late surgery based on time from symptom onset

We selected 3 different time points for early surgery: 2, 3, and 4 days from symptom onset. Twelve of 67 (21%) patients underwent surgery within 2 days of symptom onset, and 41 of 67 (61%) within 4 days or less. Demographics, tumor size, clinical symptoms, and all other features were similar when comparing patients undergoing surgery in 2 days or less vs more than 2 days from symptom onset, as well as when comparing 3 days or less vs more than 3 days and 4 days or less vs more than 4 days (see Table 7). Median tumor volume was larger in the surgical group at 2 and 3 days, but these differences were not statistically significant.

Clinical outcomes at 3 months were also unaffected by timing of surgery after symptom onset (see Table 7). No significant differences were observed in hormonal or visual function recovery comparing those who underwent surgery at 2 days or less vs more than 2 days or any other time interval evaluated, although there was a slightly better partial resolution of oculomotor palsies in the less than 2 days group.

Early vs late surgery based on time from admission or transfer to the treating center

Using a 1-day cutoff from admission/transfer until surgery, 10 of 13 (77%) patients in the early surgery group reported vision loss compared to 22 of 54 (41%) in the later surgery group (P = .019). No other presenting features correlated with early surgery (see Table 8). Using 2-day and 3-day cutoffs, the difference in vision loss disappeared, and no other findings were statistically significant.

Outcomes at 3 months were unchanged regardless of surgery timing from transfer/admission (see Table 8).

Discussion

This first-ever, multicenter, international, prospective registry of PA patients yielded 3 main findings. First, demographics, comorbidities, symptoms, and imaging findings at presentation are largely similar between patients managed surgically or medically, as are outcomes at 3 months of follow-up. Second, timing of surgery did not appear to correlate with outcomes at 3 months after surgery, suggesting there was no clear benefit to early surgery, regardless of how “early” is defined. Our findings thus clarify the mixed results seen in single-center retrospective studies, as some indicated an advantage for early surgical intervention, especially for larger pituitary adenomas (31-33), while others demonstrated essentially equal outcomes without undergoing surgery (13, 14, 20, 34-36).

We found that surgical patients had more adverse visual field defects at presentation yet similar outcomes at 3 months compared with medically managed patients. Although this might suggest that surgery led to better outcomes in patients with more severe defects, far more patients with severe defects underwent surgery and we considered that extent of visual field deficit at presentation and degree of optic nerve compression visualized on preoperative MRI are the primary features that led practitioners to choose surgical over medical management. Stratified analysis of patients with severe visual field defects or blindness did not demonstrate improved visual outcomes in the surgical group compared to the medical group at 3 or 6 months. Severe visual field deficits result from optic nerve compression, and surgical intervention can rapidly decompress neural tissues. Yet, visual field deficits of varying degrees of severity and duration are routinely observed with many large pituitary adenomas, and elective rather than urgent surgery often leads to excellent restoration of visual function (37, 38). The relatively low numbers of patients in each group limit the power of our finding but are consistent with the overall observations of each cohort. Indeed, our observation that the timing of surgical intervention had no effect on visual outcomes lends additional support to this conclusion.

Furthermore, we documented statistically significant regression in apoplectic tumor volumes at 2 to 3 months after PA onset in 18 of 19 medically managed patients, with a mean volume reduction of 66%. These finding are consistent with earlier retrospective reports. In a retrospective review of 65 patients treated over a 10-year period, of whom 49 (73%) underwent surgery and 8 (27%) were treated medically, visual and cranial nerve outcomes after 3 months or more were similar in both groups and tumor shrinkage was observed in 76.4% of those treated without surgery, suggesting that reduction in mass effect commonly occurs even in the absence of surgery (36). Similarly, in a retrospective study of 67 patients at a center that aimed to treat all patients without surgery unless they demonstrated progressive visual deterioration (defined as worsening oculomotor palsies) (39), only 17 (26.6%) had surgery, and only 7 (14.9%) deteriorated during medical treatment. At follow-up, visual and hormonal outcomes were similar in both groups, and PA volume regression of 20% or greater was seen in 95% of medically managed patients. These data indicate that tumor regression, and thus reduction in mass effect and resolution of most visual and/or or ocular defects, is a common observation in PA and occurs independent of surgical intervention.

A PAS of 3 or greater was one of the few statistically significant predictors of surgical management. As scoring largely reflects the extent of visual field defects, and nearly all patients in both groups presented with a PAS less than 4, these data further support the idea that surgical decision-making is influenced by the presence of vision loss rather than that patients with higher PAS, and thus greater visual field deficits, are more likely to see better outcomes with surgery. Importantly, we noted no patterns in either patient characteristics or outcomes according to when surgery was performed. Regardless of whether “early” surgery was defined based on time from symptom onset or time from admission to the treating center, consistent with a recent meta-analysis (19), we did not see any effect on outcome or, more important, any improved outcome, compared to patients treated later.

As we did not collect imaging results on patients at 3 months or beyond, we cannot comment on the degree of residual tumor or evidence of regrowth in patients managed medically. Nevertheless, others have shown no apparent difference in adenoma recurrence and progression rates between surgical and nonsurgical cohorts (14, 18, 20, 34, 36). Overall, adenoma recurrence or persistence was observed in 11% of patients, suggesting a need for serial follow-up scans and for decisions regarding adenomectomy based on standard criteria, including hormone levels and clinical symptoms.

We noted no differences in recovery of oculomotor function between the 2 groups, consistent with findings in retrospective series that acute decompression has no effect on this measure. In patients with pituitary adenomas and varying degrees of bitemporal hemianopia and even blindness, surgical intervention can substantially improve visual fields in 79% to 95% of patients (38, 40). By contrast, in cases of PA, our data indicate that tumor regression without intervention is likely to yield similar outcomes compared with surgical decompression.

Our registry data indicated recovery rates for anterior pituitary hormones of 6.1% to 11.6% in patients with PA, with no difference in recovery between surgical or medical groups, and no effect of timing of surgery on recovery. These rates are lower than those observed for non-PA cases, as hormonal recovery after pituitary surgery is reported to occur in 5% to 49% of patients (41, 42), with a recent multicenter analysis indicating a recovery of at least 1 axis in 21% (43), further supporting the view that surgery in patients with PA has relatively little effect on hormonal recovery.

Notwithstanding our overall results, surgery is absolutely indicated for progressive diminishment of consciousness or deteriorating visual fields despite administration of corticosteroid, as the clear benefits of acute decompression in these patients would override any other considerations. Fortunately, these situations are rare and are readily detected within the first few days of admission.

The sites that participated in this registry all have expert pituitary neurosurgery and endocrinology teams, and this may be a factor in the positive outcomes observed, especially for surgical patients. In centers without pituitary surgery expertise, a bias toward medical management may be prudent, as there are substantial data to support the view that outcomes from pituitary surgery are directly tied to surgical experience (21, 44, 45). Nevertheless, even if surgery is deemed appropriate, if the patient is clinically stable, our data indicate that, in most cases, PA is not a neurosurgical emergency.

We recognize several limitations in our study. First, there was no central review of the data. This allows for the possibility that the reported clinical, radiographic, and pathological data varied between sites, and may bias the outcomes at any given center, particularly because some measures were qualitative. For example, what one group referred to as “moderate” optic nerve compression, another group might refer to as severe. Nevertheless, the observation that both visual field findings and MRI findings of optic nerve compression were the only 2 variables found to be statistically worse in the surgical group suggests a consistency in reporting. Furthermore, our observations seem to parallel those reported in several individual institutional retrospective series regarding clinical features, MRI findings, and outcomes (1, 14, 20). As this was a registry observational study, there was no preplanned hypothesis or expected outcome, and therefore it is unlikely that bias played a role in patient selection or measurement of outcomes, and allocation for surgery vs medical management was largely consistent within each institution. Central review of all data would have eliminated any potential bias if present, but led to concerns about transferring patient information that several institutions would not allow.

A second limitation is that patients were not randomly assigned to one study group or another. This would eliminate any potential selection bias in treatment allocation, and such a study would definitively determine an independent effect of surgery on outcomes in patients with visual field deficits and/or optic nerve compression on MRI. However, in a setting where surgical intervention is an established treatment approach, IRB approval and informed consent could prove challenging, especially in an acute setting. Furthermore, we calculated that for a randomized, noninferiority trial, with a 2:1 randomization of surgery to medical management, a sample size of 357 patients (238 surgical and 119 medical) would be needed to determine a 10% difference in visual field outcomes, assuming 85% full visual fields in the surgical group, and 75% in the medical group at 3 months after presentation, based on our current data. Considering the rarity of PA and the fact that it took more than 4 years to collect data on 100 patients across 12 sites, it is unclear that such a randomized trial is feasible, even if ethically possible.

Lack of follow-up results at the 6-month time point weakens the statistical strength of our observations on patient outcomes, as does our relatively small population of only 97 patients overall and only 30 patients managed medically. Missing data points also weaken some of our conclusions. Further, as registry data are prospectively collected but not controlled, we cannot know what decisions were used at each site or in any individual case to select surgery or its timing. The large number of study sites and treatment strategies employed likely minimizes the effect of individual decisions on study results.

Despite these limitations, this study represents the largest cohort of PA patients treated at pituitary centers of excellence around the world. The results provide prospective observational data, the strongest evidence to date that, in most cases of patients presenting with PA, surgical and medical management are similarly effective, and both yield good outcomes at centers with experience in managing PA.

Conclusions

This first-ever, prospective multicenter registry of patients with PA indicates that outcomes are similar with surgical and medical management. Surgery is more often chosen as the treatment strategy in patients with visual field deficits and/or MRI evidence of optic nerve compression. However, our data do not confirm that surgery results in better outcomes for these patients, merely that it is a key factor that contributes to a decision to perform surgery. Early surgery does not appear to affect outcomes, and substantial reduction in PA tissue volumes are observed in the majority of medically managed patients within 2 to 3 months after onset. These observations suggest that in most cases, surgical and medical management will yield similarly effective outcomes in patients with PA.

Acknowledgments

The authors acknowledge the contributions of the late Dr Fred Gentili to this study and to the field of neurosurgery. We also acknowledge the help of the many data managers and study coordinators at each site without whom this project would not be possible, Jennifer Hwe for database development, and Shira Berman for editorial assistance.

Funding

None.

Disclosures

A.S.L. was a consultant for BK Medical and Spiway, and was an owner of Kogent. These relationships have concluded. G.B. is a consultant for Vascular Technologies Incorporated and Cerevasc, Inc. D.K. has received royalties from Mizuho Corporation. All other authors have nothing to declare.

Data Availability

Restrictions apply to the availability of some, or all data generated or analyzed during this study to preserve patient confidentiality. The corresponding author will on request detail the restrictions and any conditions under which access to some data may be provided.

References

Abbreviations

 
• BMI

  body mass index

 
• HIT-6

  Headache Impact Test

 
• ICU

  intensive care unit

 
• IQR

  interquartile range

 
• IRB

  institutional review board

 
• LOS

  length of stay

 
• mFI-5

  Frailty Index

 
• MRI

  magnetic resonance imaging

 
• PA

  pituitary apoplexy

 
• PAS

  Pituitary Apoplexy Score

 
• QoL

  quality of life

 
• SF-36

  Short Form Health Survey