Health-related quality of life analyses in nonfunctioning pituitary macroadenoma patients identifies at-risk populations

The quality of life (QoL) impact of multidisciplinary treatment for patients with nonfunctioning pituitary macroadenomas (NFPMA) is unclear. We sought to investigate associations between patient factors, clinical data, and patient-reported QoL in patients with NFPMA. Patients with treated NFPMA and > 1 year of follow up after transsphenoidal surgery (TSS) and with no evidence of progressive disease were evaluated utilizing the following patient-reported outcome measures: RAND-36-Item Health Survey, Multidimensional Fatigue Inventory, Cognitive Failures Questionnaire. 229 eligible patients completed QoL questionnaires a median of 7.7 years after initial transsphenoidal surgery (TSS). 25% of participants received radiation therapy (RT) a median of 2.0 years (0.1–22.5) after initial TSS. Patients who received RT were younger (median age 46 v 58, p < 0.0001), had larger tumors (28 mm v 22 mm, p < 0.0001), were more likely to have visual symptoms (65% v 34%, p = 0.0002), and were more likely to have hypopituitarism (93% v 62%, p < 0.0001). Patients with hypopituitarism reported worse energy and fatigue and cognitive function (p < 0.03). Patients who received RT reported significantly worse general health, physical health, physical fatigue and cognitive functioning (p < 0.05). The largest QoL differences were in patients who experienced a financial stressor, independent of treatment type. Hypopituitarism, radiation therapy after TSS, and financial stressors are associated with more impaired QoL in patients with NFPMA. Awareness of these factors can better guide use and timing of radiation therapy in addition to identifying patients who can benefit from multidisciplinary surveillance.


Introduction
Pituitary adenomas are commonly associated with hypersecretion of hormones, but approximately one third of all pituitary adenomas (14-54%) are nonfunctioning [1]. Clinically significant nonfunctioning pituitary macroadenomas (NFPMA) are associated with pituitary hypofunction caused by compression of the normal pituitary tissue/stalk [2] and symptoms from mass effect, including headache, vision changes, and other cranial nerve deficits. Patients with NFPMA report impaired health-related quality of life (QoL) [3], and a major goal of treatment, in addition to preventing morbidity from disease progression, is to alleviate the impaired QoL associated with the adenoma.
Depending on the clinical circumstances, management options for NFPMA include active surveillance, surgical Helen A. Shih hshih@mgh.harvard.edu resection, and radiation therapy (RT). Transsphenoidal surgery (TSS) is the preferred treatment for large, growing, or symptomatic tumors, but 34% of patients with radiographic complete resection and 70% of patients with incomplete resection after TSS experience tumor regrowth by 15 years after surgery [4]. While there are no randomized controlled trials to guide selection of patients who receive RT after TSS, RT is usually reserved for incompletely resected, recurrent, or progressive NFPMA [2]. Unfortunately, treatment with TSS and RT can cause complications that further impair QoL.
When quantifying the clinical impacts of NFPMA, physicians typically focus on tumor size, hormone excess or deficiency due to compression of the normal pituitary tissue and/or hyperprolactinemia from compression of the stalk, and the neurologic symptoms associated with the tumor and its treatment. However, previous radiation therapy trials in other diseases (lung and prostate cancer) have demonstrated that, for subjective symptoms, patient-reported outcomes are more reflective of the patient experience than symptom ratings by medical professionals [5,6]. For some cancer types, patient-reported QoL is even associated with survival [7,8].
The degree to which QoL is impaired, the QoL domains affected, and the mechanisms of QoL impairment after treatment of NFPMA are unclear. Additionally, the effect of RT on QoL in patients with NFPMA is uncertain, as some studies suggest that prior RT is associated with impaired QoL [9,10], while others suggest that it is not [11][12][13].
This cross-sectional study of patients at least one year after initial TSS for NFPMA is the largest study of QoL in patients with NFPMA of our knowledge to date. We sought to investigate (1) whether QoL is impaired in patients with NFPMA, and if so, in what domains, (2) whether RT for residual or recurrent disease is associated with decreased QoL, and (3) what treatment, patient, and clinical factors are associated with decreased QoL. Results of this study identify patient populations at higher risk for decreased QoL after treatment for NFPMA who may benefit from increased surveillance and early intervention.

Eligibility criteria and recruitment of participants
Eligible participants were English-speaking adult patients (> 18 years old) with nonfunctioning pituitary macroadenomas (> 1 cm) whose initial treatment consisted of transsphenoidal surgery (TSS) between 1980 and 2010. Some patients received subsequent treatment after initial TSS, including additional surgery or radiation therapy for residual or recurrent disease. All eligible patients had at least one year of follow up from their most recent treatment and were free of any evidence of macroadenoma progression on their most recent imaging. Participants meeting these criteria were identified by querying patient lists from the Massachusetts General Hospital (MGH) Neuroendocrine and Pituitary Tumor Clinical Center and the MGB Patient Data Registry. Electronic medical records were reviewed to determine preliminary eligibility. Patients who passed this screen were contacted by mail with a recruitment letter, study description sheet, and study questionnaires (described below). Patients who wished to participate gave written informed consent and returned the questionnaires in a prepaid envelope. Study staff attempted to contact patients by telephone if they did not return the questionnaires.

Treatment and follow up of patients
All eligible patients received TSS as initial treatment of their NFPMA. The need for subsequent treatment was determined by managing clinicians. Follow up for these patients generally consisted of yearly history and physical exam, laboratory testing, and MRI. Many, but not all, patients were followed at the MGH Neuroendocrine and Pituitary Tumor Clinical Center. All patients received some portion of their treatment or follow up at MGH.

Measures
The questionnaires consisted of validated patient reported outcome measures (the RAND 36-Item Health Survey, version 1 (RAND-36), the Multidimensional Fatigue Inventory , and the Cognitive Failures Questionnaire (CFQ)), and one study-specific questionnaire to collect additional data on demographics, pituitary adenoma treatment, comorbidities, and follow up.
The RAND-36 questionnaire v1.0 is a self-assessment that evaluates 8 dimensions (physical functioning; role limitations due to physical problems; pain; general health perceptions; emotional well-being; role limitations due to emotional problems; energy/fatigue; and social functioning) and 3 composite domains (physical health, mental health, and global health). Scoring was performed as previously described [14,15] and scores were compared to 1990 ageand sex-adjusted general population norms [16]. The 3 composite domains are T-scores adjusted to population norms, where a difference of 10 is one standard deviation of the general population scores. Differences of at least 3-5 points are considered clinically meaningful [14,17].
The CFQ is a 25-item questionnaire that measures selfreported frequency of daily cognitive errors over the prior 6 months [20]. The summary score (0-100, with higher scores representing worse QoL) within our patient population was calculated. Differences of at least 3 points are considered clinically meaningful [21,22].
All necessary permissions were obtained to use the QoL questionnaires. Missing data for RAND-36 were addressed as previously described [23]. Any MFI-20 subscales or CFQ questionnaires with missing data from individual questions were omitted. Primary outcome measures were the physical and mental health composite T-scores of the RAND-36, the MFI-20 general fatigue subscale score, and the CFQ total score. Secondary outcome measures were the additional subscales of the RAND-36 and the MFI-20. A significance level of 0.05 was used for all analyses without adjustment for multiple comparisons due to the exploratory nature of the analyses. SAS v9.4 software (SAS Institute, Inc., Cary, NC) was used for the analyses.
Free text responses to questions regarding recent events or situations that might have impacted responses to the QOL surveys were used to define yes/no variables reflecting current medical, financial, or social/cognitive stressors.

General self-report questionnaire and chart review
We collected additional data on demographics, disease severity, presentation, NFPMA treatment, comorbidities, and follow up via a self-reported questionnaire that we mailed to patients along with the QoL questionnaires and via review of patients' medical records. In terms of baseline (at time of initial TSS) characteristics, we collected the following by chart review: whether NFPMA was diagnosed incidentally or due to symptoms, year of TSS, age at TSS, greatest tumor diameter on MRI, presence of visual symptoms, presence of apoplexy, presence of hormone-related symptoms at TSS, evidence of hypopituitarism, and Charlson comorbidity score. We collected the following additional data regarding treatment of NFPMA by chart review and self-report: existence and year of any subsequent treatments (surgery, radiation therapy, or medications such as dopamine agonists) after initial TSS, replaced and unreplaced pituitary hormone deficiencies (thyroid, glucocorticoid, growth hormone, sex hormone (in women under 50 and in men regardless of age), and antidiuretic hormone (ADH)), Charlson comorbidity score, date of most recent follow up at MGH, self-reported medications, and for patients treated with RT, presence of hypopituitarism at the time of RT, RT modality (photon vs. proton and traditionally fractionated vs. single-fraction stereotactic radiosurgery), RT dose, and RT hospital (MGH vs. other). Hypopituitarism was defined as either a documented deficiency of thyroid hormone, glucocorticoids, sex hormones (in women under 50 and in men regardless of age), or growth hormone, or a prescription for replacement of any of these hormones. We collected the following demographic data by chart review and self-report at the time of QoL survey: sex, marital status, living situation, education level, self-reported financial situation (very secure/comfortable, reasonably secure, unstable/insecure, or very dire), and selfreported recent life event(s)/stressor(s) affecting perceived QoL. Missing data were recorded as such. No patient was considered to have missing data with regard to Charlson comorbidity score. Self-reported illnesses and medications were used to assign comorbidities to patients who did not have recent follow up at MGH.

Statistical analysis
Characteristics of QoL survey responders and non-responders were compared using Fisher's exact tests and Wilcoxon rank-sum tests. For responders, baseline (at time of TSS) and current (at the time of QoL assessment) characteristics of those who never received RT ("TSS-only" group) were compared to those who subsequently received RT at some point after initial TSS ("TSS + RT" group) using Fisher's exact tests and Wilcoxon rank-sum tests. For all 8 RAND-36 subscales, z-scores reflecting standardized differences in QoL score between study population and age-and sexadjusted general population means were calculated to compare QoL in the study population to an appropriate control group, in both the overall sample and within the TSS-only and TSS + RT groups. Mean z-scores were compared to zero using a one-sample t-test. Mean scores for all RAND-36 subscales and composites, for the MFI-20 subscales, and for CFQ scale were also compared between TSS-only and TSS + RT groups using univariate linear regression. Because of differences in follow up time between these two groups, a subgroup analysis was performed to compare QoL scores between the TSS-only and TSS + RT groups in the subset of patients with at least 5 years of follow up since initial TSS.
Associations between other characteristics and QoL scores were compared using univariate linear regression. Factors associated with RT or with QoL on any of the primary outcome measures were entered into a multivariate regression for each primary and secondary QoL outcome to create the final multivariate models.
Covariates that were significant in the model(s) for one or more of the 3 primary outcome measures (i.e. the Physical and Mental composites of the RAND-36, the MFI-20 General Fatigue subscale, and the CFQ total score) were years earlier than the TSS-only group (median year of initial TSS 1997 vs. 2005, p < 0.001), and therefore has longer follow up from time of initial treatment to QoL assessment (median 14.0 vs. 6.3 years). Table 2 summarizes clinical and treatment information for the 58 patients treated with radiation therapy. 83% received RT at MGH, while the remaining 10 received RT elsewhere. RT modality information was missing for 9 patients, and of the other 49 patients, 25 (51%) were treated with fractionated photon RT, 10 (20%) were treated with fractionated proton RT, and 14 (29%) were treated with single-fraction proton stereotactic radiosurgery (SRS).
Sixteen patients (28%) received radiation treatment within one year of initial TSS, 22 (38%) were treated more than a year after initial TSS but before any subsequent surgery, and the remaining 20 patients (34%) were treated with radiation therapy after a second or third surgery ( Table 2). Nine of 58 patients (16%) required at least one additional surgery at a median of 10.4 years after completion of radiation therapy. 35% of patients who eventually received radiation therapy had hypopituitarism at the time of initial TSS (Table 1), while 66% of these patients had evidence of hypopituitarism by the time they received RT ( Table 2). Table 3 summarizes patient characteristics and clinical features at the time of quality of life survey completion by treatment group. 61% of TSS-only patients and 91% of TSS + RT patients had at least 5 years of follow up between initial TSS and QoL assessment (p < 0.001). The TSS-only group was older than the TSS + RT group at the time of QoL assessment (median 65 vs. 60 years, p = 0.011), but marital status, living situation, and education level were not significantly different between groups ( Table 3). RT was associated with self-reported unstable/insecure or very dire financial situation (28% vs. 7%, p < 0.001), as well as selfreported current financial stressor (16% vs. 5%, p = 0.016).
At the time of the QoL survey, patients who received RT after TSS were more likely to have hypopituitarism (93% vs. 62%, p < 0.001) and more likely to be on hormone replacement treatment for all pituitary hormones except ADH (Table 3). Because hypopituitarism is a late effect of radiation treatment, we specifically examined the subset of patients with at least 5 years of follow up. In these patients, the total rates of replaced or replaced deficiencies in thyroid, glucocorticoid, growth hormone, and sex hormone were 83%, 62%, 26%, and 69%, respectively, for the TSS + RT group vs. 41%, 24%, 9%, and 44%, respectively, for the TSS-only group.
Compared to patients who had TSS alone, those who received radiation therapy reported worse QoL scores across numerous domains using multiple QoL assessment tools. Patients who received RT reported significantly worse QoL in terms of physical limitations (76.5 vs. 64.9, identified via backwards selection. These identified covariates were then used to create multivariate models for each of the QoL/utility outcomes, with each model containing the same covariates. We excluded the "incidental diagnosis" covariate from this analysis due to missing data. We included the covariates of at least 1 additional surgery after initial TSS and follow-up time regardless of significance because of their strong association with RT. The final 7 covariates that were included in the 17 multivariate models are as follows: RT, additional surgery after initial TSS, at least 5 years of follow up from initial TSS to QoL survey, age at QoL survey, Charlson score at QoL survey, hypopituitarism at QoL survey, and self-reported financial situation/ stressor at QoL survey.

Results
There were 488 potentially eligible patients who received TSS for NFPMA at Massachusetts General Hospital (MGH) between 1982 and 2010. Recruitment letters and QoL questionnaires were sent to these patients and 244 eligible patients (54%) returned the completed questionnaires. Fifteen patients were found to be ineligible upon further review of medical records, and the remaining 229 patients were included in the analyses below.
Non-responders were younger at initial TSS than responders (median 50 vs. 55 years, p = 0.0048) and had TSS earlier (median year of TSS 2002 vs. 2004, p = 0.0327) than responders. Non-responders and responders did not differ significantly with respect to sex (p = 0.2091), receipt of RT (p > 0.99), or size of NFPMA prior to TSS (p = 0.4295). Table 1 summarizes baseline patient characteristics, baseline clinical features, treatment details, and follow up details for patients included in the analysis by treatment group. The median year of initial TSS for all 229 participants was 2004. 35% were women and 65% were men with a median age of 55 years at the time of initial TSS. Fifty-eight participants (25%) received radiation therapy a median of 2 years after initial TSS, and 35 patients (15%) had at least one additional surgery after initial TSS.
The patients who received RT were younger (median age 46 vs. 58 years, p < 0.001), had larger tumors at initial TSS (28 mm vs. 22 mm, p < 0.001), and were more likely to have visual symptoms before TSS (65% vs. 34%, p = 0.002) ( Table 1). There were no statistically significant differences in sex, hormone-related symptoms, hormonal deficiencies, or pituitary apoplexy between patients treated with TSS and TSS + RT. While only 5% of patients who never received radiation therapy had at least one surgery after initial TSS, 47% of those who did receive RT required additional surgery (p < 0.001). The TSS + RT group had initial TSS 8 group. Beyond whether patients received RT, QoL was also significantly associated with a number of other covariates on univariate (Supplemental Table S1) and multivariate analyses (Table 5). Interestingly, patients who were older than 65 at the time of QoL survey completion reported higher QoL for physical functioning, general health perceptions, energy level, emotional well-being, social functioning, and mental health.
Patients with hypopituitarism (corrected or uncorrected) at completion of their QoL surveys reported worse (RAND-36 Energy/Fatigue: -7.95, p = 0.026) and cognitive function (CFQ: 5.35, p = 0.026). Interestingly, the largest QoL differences were seen in patients who experienced a financial stressor after treatment for their NFPMA. These patients  (Table 4). Interestingly, in the subset of patients with at least 5 years of follow up, the differences between patients who received TSS vs. TSS + RT decreased, with only the difference in RAND-36 Energy subscale score remaining statistically significant (60.6 vs. 51.9, p = 0.043). Table 5 summarizes a multivariate analysis between QOL outcomes and patient characteristics and treatments. After adjusting for other covariates, patients who received radiation therapy had worse QoL in terms of general health (-8.44, p < 0.05), physical health composite score (-4.08, p < 0.05), physical fatigue (11.68, p < 0.05), and cognitive functioning (6.64, p < 0.05) compared to the TSS-only  At Least 5 Years Follow Up from Initial TSS (n, %) 158 (69%) 105 (61%) 53 (91%) < 0.001 0 (0%) / 0 (0%) a Number of patients with non-missing data varied from 175 to 229 for the overall cohort, from 141 to 171 for the TSS group, and from 34 to 58 for the TSS-with-subsequent-RT group, depending on the individual variable. For categorical variables, percentages are for available data. b Of those patients with hormone deficiencies, 63% (64/102), 63% (49/78), and 63% (15/24) in each group, respectively, were taking medication to correct the deficiency prior to TSS. c Of patients with TSH, ADH, GH, and/or ACTH deficiencies, 85% (58/68), 83% (45/54), and 93% (13/14) in each group, respectively, were taking medication to correct the deficiency prior to TSS. d Applies only to patients who had a subsequent treatment (i.e. additional surgery, RT, or dopamine agonist medication) after initial TSS.

Discussion
With 229 analyzable patients, this cross-sectional study of quality of life in patients with NFPMA is to our knowledge the largest study to date of health-related QoL in this population. MGH is a quaternary care referral center with surgical, medical, and radiation expertise in the management of pituitary tumors. Our radiation expertise includes early adoption of advanced radiation technologies for the management of pituitary tumors, including stereotactic radiosurgery and proton radiotherapy. Therefore, our patients likely have more advanced disease than the average NFPMA patient, which is likely what explains the lower local control rates after RT than typically reported; with a median of 7.1 years of follow up after RT, 15.5% of patients required at least one additional surgery. However, patients who had radiation therapy and required subsequent surgery had a 10.4 outcomes, including emotional well-being, physical and mental health, social functioning, energy level, and motivation (Table 5).
Compared to population norms, patients who received TSS but not RT reported significantly better QoL in terms of physical functioning, physical limitations, pain, and general health perceptions (Supplemental Table S2), possibly a reflection of the overall better health of patients who are surgical candidates. In contrast, these domains were not significantly different from population norms for patients who received TSS + RT, but energy and fatigue were significantly worse in this group compared to population norms (Supplemental Table S2). patients in whom long-term local control is a priority. Indeed, this series demonstrates that patients who received radiation treatment were younger, had larger tumors, and were more likely to present with symptoms (including visual symptoms), rather than smaller, incidentally diagnosed tumors. These patients are at higher lifetime risk for local recurrence, and were therefore more likely to require year (0.8-23.6) median interval between adjuvant radiation therapy and subsequent surgery, suggesting that upfront adjuvant RT increases the durability of tumor control.
While there are no randomized phase 3 clinical trials to guide the use of radiation therapy in NFPMA, postoperative radiation therapy is typically used for patients with more advanced, aggressive or invasive disease and in younger associated with cognitive functioning [24]. However, our sample size of 58 patients who received RT and responded to QoL surveys was not powered to assess whether radiation treatment modality impacts patient-reported QoL. The relationship between radiation therapy, hypopituitarism, and quality of life is complex. Patients with a larger baseline tumor volume are more likely to develop secondary hypopituitarism [25] and are more likely to receive radiation treatment, which itself is independently associated with hypopituitarism [26,27]. In this study, patients who received RT after TSS had higher rates of hypopituitarism, multiple surgeries and adjuvant radiation therapy. Patients who received radiation therapy generally reported poorer QoL in terms of general health, physical health, fatigue, and cognitive function, even after controlling for hypopituitarism. The cause of these QoL impairments is likely multifactorial, but may be impacted by the fact that these patients had more advanced disease, multiple surgeries, and were younger at the time of treatment. Changes in cognitive function may also be related to irradiation of normal tissues. Proton therapy, which allows for more sparing of normal tissue, may be one approach to reduce the decrement in QoL their financial situation was either "unstable/insecure" or "very dire". The phenomenon of financial toxicity associated with cancer treatment is well-documented [28], but is also applicable to treatment of benign conditions such as NFPMA. In addition to the cost of medical interventions, it is also possible that the time commitment required for fractionated radiation therapy or the associated decreases in fatigue or cognitive function affect patients' job security and financial status after radiation treatment. Financial toxicity is frequently associated with socioeconomic status and social support [29,30], but the sociodemographic and clinical features that may increase the risk for financial consequences of treatment in patients with NFPMA are unknown. Our study suggests that patients who report poor financial circumstances at the time of treatment for their NFPMA and patients receiving radiation therapy may be at an increased risk of having poor QoL after treatment. Incorporating social workers or other psychosocial support services into the multidisciplinary care of patients with pituitary tumors may allow early identification of the patients most in need of intervention and thus may result in better quality of life for these patients. Although not performed in this study, future studies should examine the correlation between socioeconomic status and self-reported financial stressors to further improve early identification of patients who may report poor QoL after treatment of NFPMA. both at baseline and at the time of QoL survey, likely a function of disease severity, tumor size, and treatment effect. In patients with more than 5 years of follow up, hypopituitarism was independently associated with worse QoL in terms of energy/fatigue and cognitive functioning. Given that patients who received RT were more likely to have a greater total number of hormone deficiencies, it is also possible that corrected hormone deficiencies resulted in fatigue due to the absence of physiologic fluctuations with hormone replacement. Given the high rate of pituitary dysfunction after radiation therapy and the time interval between completion of radiation therapy and hormonal dysfunction, patients should be counseled at the time of radiation therapy and at subsequent follow up visits about the importance of neuroendocrine evaluation by an endocrinologist with expertise in pituitary disorders and surveillance for development of hypopituitarism to prevent a hypopituitarism-associated diminution in QoL as well as serious hypopituitarism-associated medical complications, including adrenal crisis and hyponatremia.
In this study, the largest decreases in patient-reported QoL were seen in patients with a self-reported financial stressor. Patients who reported a financial stressor had worse average QoL scores in all domains examined, and this independent association was both statistically and clinically significant in 10 of 17 domains (Table 5). Furthermore, patients who received radiation therapy were more likely to report that The level of significance (p value) from multivariate Wald tests of the estimated coefficients is indicated by the number of asterisks (*0.05 ≤ p < 0.1; **0.01 ≤ p < 0.05; ***p < 0.01).
One limitation of this study, which affects many QoL studies, is the issue of patient response rates. While the questionnaire completion rate (54%) is high for a QoL study [31,32], perhaps indicative of patient satisfaction of their care with our providers, it is possible that there are nonrandom factors underlying the missing QoL data. Furthermore, this cross-sectional study does not account for baseline QoL or changes over time, and uses population norms for comparison, rather than an age-and sex-matched cohort for comparison. Other limitations of this cross-sectional study include differences in the year of surgery and duration of follow up between the patients who required TSS alone vs. TSS and radiation therapy. These differences reflect the fact that patients who required radiation therapy after TSS were more likely to require multiple interventions. Therefore, these patients had longer interactions with the healthcare system and thus longer follow up data after their initial TSS. The mean patient age at time of initial TSS also differs between the TSS and TSS + RT groups, which reflects the fact that younger NFPMA patients are more likely to be treated with both surgery and radiation therapy, because the likelihood of recurrence increases over time.
Importantly, this study emphasizes the importance of the timing of radiation therapy after surgical resection of NFPMA. There are variations in practice patterns, particularly in the setting of subtotal resection, with some radiation oncologists recommending adjuvant radiation therapy immediately after subtotal resection, while others recommend adjuvant treatment only after radiographic progression. In one study of 663 patients who underwent surgery for NFPMAs, only 45% of patients with a subtotal resection and no adjuvant radiation therapy had disease recurrence [33]. There was no difference in survival between patients who received radiotherapy adjuvantly or at the time of progression. These data, in combination with our findings regarding the effects of radiotherapy on QoL, suggest that waiting for radiographic progression before treating with adjuvant radiotherapy can preserve patient QoL and prevent overtreatment without affecting survival outcomes.
In conclusion, this QoL analysis of patients with nonfunctioning pituitary macroadenomas demonstrates that pretreatment patient characteristics, treatment with radiation therapy, hypopituitarism, and financial stressors are associated with significantly decreased QoL across numerous domains. Long-term follow up for these patients is essential, and early establishment of a multidisciplinary team including endocrinology and psychosocial support may prevent disease-and treatment-related decreases in patient quality of life.