Keywords
pituitary tumor - transsphenoidal surgery - health care disparity - practice setting
Introduction
The treatment of patients with pituitary tumors is complex and multidisciplinary,
often involving surgical and medical treatments as well as the potential for radiotherapy.
Protocols for the surgical management of these tumors by otolaryngologists and neurosurgeons
have been well characterized in the literature.[1]
[2]
[3] At the same time, little is known about whether socioeconomic disparities exist
in the context of pituitary surgery. Characterizing these disparities is essential
to ensuring that all patients receive the highest possible standard of care. Efforts
to identify and alleviate socioeconomic health care disparities have taken place over
several decades, such as the creation of the World Health Organization Commission
on Social Determinants of Health in 2005.[4] More recently, the importance of addressing socioeconomic health care disparities
was highlighted by the current coronavirus disease 2019 (COVID-19) pandemic, which
disproportionately affected minority groups such as African Americans and Native Americans.[5]
While the COVID-19 pandemic remains a salient example of socioeconomic disparities
in health care, numerous examples of similar disparities have been described in the
surgical literature. Several studies have identified racial and socioeconomic disparities
in surgical specialties including general surgery,[6] surgical oncology,[7]
[8] cardiothoracic surgery,[9] and orthopaedic surgery.[10] In spite of this, relatively little work has attempted to uncover these disparities
in otolaryngology and neurosurgery. Some studies have characterized disparities in
conditions such as sinusitis[11]
[12]
[13] and squamous cell carcinoma of the head and neck.[14]
[15] Other studies in neurosurgery have demonstrated poorer outcomes after ischemic strokes
among patients without private insurance[16] and greater complication rates among African American patients undergoing craniotomies
for glioma resection.[17]
One scoping review of neurosurgical research found that the majority of socioeconomic
disparities research in neurosurgery is concerned with the role of race and income
in surgical outcomes.[18] Furthermore, spine surgery, vascular neurosurgery, and cranial oncology have been
studied most extensively in the context of socioeconomic disparities. At the same
time, there is limited evidence on the impact of health care disparities in the context
of particular pathologic entities within these categories. At present, few studies
have addressed racial and socioeconomic disparities in the context of pituitary surgery.[19]
[20]
[21] Most studies have employed large databases such as the National Inpatient Sample
and the New York Statewide Planning and Research Cooperative System to determine the
association between complication rates and socioeconomic variables such as race, income,
and insurance status. They have demonstrated higher complication rates and greater
postoperative length of stay among black or Hispanic patients.[19]
[20] Moreover, these studies have suggested that patients belonging to racial minority
groups or who have Medicaid are more likely to be treated at low-volume centers.[19]
[21] However, studies utilizing large databases have significant limitations such as
missing data and selection bias. Other studies describing single-center or multicenter
cohorts have corroborated these findings, demonstrating that patients in racial minority
groups are more likely to present with advanced disease and experience postoperative
complications.[22]
[23]
The close relationship between NYU Langone Health and Bellevue Hospital presents a
unique opportunity to further our understanding of disparities in pituitary surgery.
NYU Langone Health is a leading academic medical center and is the home of the NYU
Grossman School of Medicine. In contrast, Bellevue Hospital is the oldest public hospital
in the United States, and it serves as the main tertiary care center for the public
hospital system of New York City. Despite being located three blocks from one another,
these institutions differ considerably in the socioeconomic composition of their patient
populations; however, they share common faculty members and residents. Accordingly,
comparing the management of pituitary tumors between these hospitals would allow us
to assess whether socioeconomic and racial disparities exist in pituitary surgery
while controlling for hospital faculty and house staff.
Methods
The Institutional Review Boards of the New York University Grossman School of Medicine
and Bellevue Hospital approved this retrospective chart review. Inclusion criteria
were patients aged 18 to 90 who underwent endoscopic endonasal pituitary adenoma resection
at NYU Langone Health or Bellevue Hospital between October 2011 and July 2020. This
timeframe was chosen due to the availability of information from the medical record
systems of each hospital. Patients with nonadenoma pituitary tumors and those without
pathology reports were excluded. There were 72 Bellevue patients that met inclusion
criteria within the time frame. Because NYU Langone Health has significantly larger
case volume, consecutive patients starting July 2020 (and working retrospectively)
were screened for inclusion until 72 subjects were identified for that group. Baseline
socioeconomic variables collected included patient-identified race and ethnicity,
primary payer, ability to speak English, and domiciled status. Race was defined as
American Indian, Asian, Hispanic, non-Hispanic black, and non-Hispanic white. Medical
history of diabetes mellitus, hypertension, obesity, and smoking were recorded. When
available, tumor diameter and preoperative endocrine abnormalities were also recorded.
Lastly, we recorded whether a preoperative eye exam was documented and whether a visual
field deficit (unilateral or bilateral hemianopia) was observed.
Patients were grouped by the hospital where they underwent surgery. Subgroup analyses
were then performed on the entire cohort based on racial/ethnic minority status and
insurance type (private vs. other). Intraoperative variables including procedural
duration and development of intraoperative cerebrospinal fluid (CSF) leak were recorded.
The primary outcome was time to surgery from the initial neurosurgical recommendation,
excluding patients who presented for emergency surgery. Emergency surgery was defined
as procedures performed during the first presentation and admission at either institution.
Reasons for emergency surgery included pituitary apoplexy, cranial nerve deficits,
or sudden visual impairment. Secondary outcomes included symptom duration, intraoperative
CSF leak, length of stay, gross total resection rate, postoperative CSF leak, postoperative
diabetes insipidus, and hormone supplementation.
Statistical Analysis
Categorical variables were assessed using chi-square or Fisher's exact tests as appropriate.
Continuous variables were assessed for normality using the Shapiro–Wilk test. Subsequently,
continuous variables were analyzed with two-sample t-tests and Mann–Whitney U tests as appropriate. All analyses were performed using SPSS 28.0 (IBM, Armonk, New
York, United States). Significance testing was two-sided with a 5% α level.
Results
A total of 144 patients were included in this study, 72 from NYU Langone Health and
72 from Bellevue Hospital. The average age was 48.8 years (standard error: 1.2), 85
patients (56.7%) were female, and the most common racial/ethnic group reported was
Hispanic (46, 31.9%). Most patients (97, 67.4%) presented with a nonfunctioning adenoma.
Among patients with functional adenomas, growth hormone-secreting adenomas were the
most common type (19, 40.4%).
On average, private hospital patients were 5.6 years older than public hospital patients
(p = 0.06, [Table 1]). Compared with private hospital patients, public hospital patients were more likely
to identify as a racial/ethnic minority group (87.5% vs. 61.1%, p < 0.001) and were less likely to have private insurance (8.3% vs. 76.4%, p < 0.001) or speak English (43.1% vs. 88.9%, p < 0.001). No significant differences between patients of both hospitals were observed
for medical comorbidities, smoking status, symptom duration, or preoperative endocrine
abnormalities. Similarly, no differences were seen in history of previous transsphenoidal
surgery, tumor diameter, or whether a preoperative eye exam was performed; however,
patients at Bellevue Hospital were more likely to have a documented visual field deficit
(52.8% vs. 30.6%, p = 0.01).and were more likely to require emergency surgery (12.5% vs. 2.8%, p = 0.03). By contrast, patients at NYU Langone Health were more likely to present
with cavernous sinus invasion on preoperative imaging (48.6% vs. 29.2%, p = 0.04).
Table 1
Comparisons between patients treated at a private and public hospital
|
Public (n = 72)
|
Private (n = 72)
|
p-Value
|
|
Age (y)
|
46.5 (1.6)
|
51.1 (1.9)
|
0.06
|
|
Sex (female)
|
43 (59.7%)
|
40 (55.6%)
|
0.61
|
|
Race/Ethnicity
|
|
American Indian
|
1 (1.4%)
|
0 (0.0%)
|
< 0.001[a]
|
|
Asian
|
13 (18.1%)
|
5 (6.9%)
|
|
Hispanic
|
34 (47.2%)
|
12 (16.7%)
|
|
NH-black
|
15 (20.8%)
|
27 (37.5%)
|
|
NH-white
|
6 (8.3%)
|
24 (33.3%)
|
|
Other/Unknown
|
3 (4.2%)
|
4 (5.6%)
|
|
English-speaking
|
31 (43.1%)
|
64 (88.9%)
|
< 0.001[a]
|
|
Undomiciled
|
2 (2.8%)
|
0 (0.0%)
|
0.24
|
|
Primary payer
|
|
Medicaid
|
33 (45.8%)
|
4 (5.6%)
|
< 0.001[a]
|
|
Medicare
|
9 (12.5%)
|
13 (18.1%)
|
|
Private insurance
|
6 (8.3%)
|
55 (76.4%)
|
|
Uninsured
|
14 (19.4%)
|
0 (0.0%)
|
|
Unknown
|
10 (13.9%)
|
0 (0.0%)
|
|
Diabetes mellitus
|
11 (15.3%)
|
13 (18.1%)
|
0.66
|
|
Hypertension
|
16 (22.2%)
|
27 (37.5%)
|
0.05
|
|
Obesity
|
9 (12.5%)
|
9 (12.5%)
|
0.99
|
|
Current smoker
|
6 (8.3%)
|
5 (6.9%)
|
0.77
|
|
Tumor type
|
|
Nonfunctioning adenoma
|
49 (68.1%)
|
48 (66.7%)
|
0.29
|
|
ACTH adenoma
|
3 (4.2%)
|
8 (11.1%)
|
|
FSH adenoma
|
1 (1.4%)
|
1 (1.4%)
|
|
GH adenoma
|
13 (18.1%)
|
6 (8.3%)
|
|
Prolactinoma
|
6 (8.3%)
|
8 (11.1%)
|
|
TSH adenoma
|
0 (0.0%)
|
1 (1.4%)
|
|
Symptom duration (mo)
|
15.2 (2.7)
|
13.2 (1.4)
|
0.28
|
|
Emergency surgery
|
9 (12.5%)
|
2 (2.8%)
|
0.03[a]
|
|
Preoperative hormone abnormalities
|
|
Deficiency
|
30 (41.7%)
|
26 (36.1%)
|
0.68
|
|
Excess
|
41 (56.9%)
|
41 (56.9%)
|
0.67
|
|
Any
|
54 (75.0%)
|
51 (70.8%)
|
0.99
|
|
Previous surgery
|
8 (11.1%)
|
13 (18.1%)
|
0.24
|
|
Time to surgery (d)
|
46.2 (7.8)
|
34.8 (3.7)
|
0.39
|
|
Preoperative eye exam
|
66 (91.7%)
|
61 (84.7%)
|
0.20
|
|
Visual field deficit
|
38 (52.8%)
|
22 (30.6%)
|
0.01[a]
|
|
Tumor size (mm)
|
28.9 (1.7)
|
28.1 (1.9)
|
0.65
|
|
Cavernous sinus invasion
|
21 (29.2%)
|
35 (48.6%)
|
0.04[a]
|
|
Procedure length (min)
|
254.3 (11.0)
|
268.9 (15.4)
|
0.96
|
|
Intraoperative CSF leak
|
36 (50.0%)
|
39 (54.2%)
|
0.62
|
|
Length of stay (days)
|
8.5 (1.9)
|
5.5 (0.4)
|
0.05
|
|
Postoperative CSF leak
|
7 (9.7%)
|
11 (14.7%)
|
0.31
|
|
Transient diabetes insipidus
|
26 (36.1%)
|
13 (18.1%)
|
0.02[a]
|
|
Postoperative DDAVP supplementation
|
5 (6.9%)
|
12 (16.7%)
|
0.08
|
|
Reoperation
|
7 (9.7%)
|
8 (11.1%)
|
0.76
|
|
Readmission
|
6 (8.3%)
|
12 (16.7%)
|
0.12
|
|
Gross total resection
|
38 (52.8%)
|
50 (69.4%)
|
0.04[a]
|
|
At least one follow-up
|
69 (95.8%)
|
70 (97.2%)
|
0.65
|
Abbreviations: ACTH, adrenocorticotropic hormone; CSF, cerebrospinal fluid; DDAVP,
desmopressin; FSH, follicle-stimulating hormone; GH, growth hormone; NH, non-Hispanic;
TSH, thyroid stimulating hormone.
Note: Continuous variables reported as mean with standard error.
a Statistical significance.
Though 11 patients (7.6%) who presented for emergency surgery were included in this
study, they were excluded when comparing time-to-surgery between the two facilities.
Among the remaining 133 patients (92.4%), the average time to surgery from initial
neurosurgical recommendation was 11.4 days lower at NYU Langone Health than at Bellevue
Hospital, though this finding was not statistically significant (p = 0.39). No significant differences were seen in procedure length or development
of intraoperative CSF leaks. Bellevue Hospital patients were more likely to develop
transient diabetes insipidus (p = 0.02). No differences were seen in rates of postoperative CSF leaks, length of
stay, 30-day reoperation, 30-day readmission, presentation for follow-up, or use of
desmopressin supplementation. Based on postoperative imaging, gross total resection
was more commonly achieved in private hospital patients than public hospital patients
(69.4% vs. 52.8%, p = 0.04).
When stratifying patients of both hospitals by racial/ethnic minority status, non-Hispanic
white patients had surgery 15.9 days earlier than patients belonging to minority groups,
though this was not statistically significant ([Table 2]). Minority group patients were more likely to develop diabetes insipidus (p = 0.04). A smaller difference in time-to-surgery (7.7 days) was observed when comparing
patients with private insurance and those with other insurance types (p = 0.99, [Table 3]). In addition, patients with private insurance were more likely to have gross total
resection achieved (p = 0.02).
Table 2
Comparisons of outcomes based on racial/ethnic minority status
|
Nonminority (n = 30)
|
Minority (n = 107)
|
p-Value
|
|
Time to surgery (d)
|
28.5 (3.7)
|
44.4 (5.8)
|
0.19
|
|
Length of stay (d)
|
6.0 (0.9)
|
7.4 (1.3)
|
0.33
|
|
Intraoperative CSF leak
|
18 (60.0%)
|
54 (50.5%)
|
0.36
|
|
Postoperative CSF leak
|
6 (20.0%)
|
12 (11.2%)
|
0.21
|
|
Transient diabetes insipidus
|
4 (13.3%)
|
35 (32.7%)
|
0.04[a]
|
|
Postoperative DDAVP supplementation
|
3 (10.0%)
|
14 (13.1%)
|
0.76
|
|
Reoperation
|
4 (13.3%)
|
11 (10.3%)
|
0.74
|
|
Readmission
|
1 (3.3%)
|
17 (15.9%)
|
0.12
|
|
Gross total resection
|
21 (70.0%)
|
64 (59.8%)
|
0.31
|
|
At least one follow-up
|
30 (100.0%)
|
102 (95.3%)
|
0.59
|
Abbreviations: CSF, cerebrospinal fluid; DDAVP, desmopressin.
Note: Continuous variables reported as mean with standard error.
a Statistical significance.
Table 3
Comparisons of outcomes based on insurance status
|
Private insurance (n = 61)
|
Other insurance (n = 73)
|
p-Value
|
|
Time to surgery (d)
|
36.9 (4.1)
|
44.6 (7.5)
|
0.99
|
|
Length of stay (d)
|
5.7 (0.5)
|
8.2 (1.9)
|
0.30
|
|
Intraoperative CSF leak
|
33 (54.1%)
|
36 (49.3%)
|
0.20
|
|
Postoperative CSF leak
|
10 (16.4%)
|
8 (11.0%)
|
0.36
|
|
Transient diabetes insipidus
|
12 (19.7%)
|
22 (30.1%)
|
0.17
|
|
Postoperative DDAVP supplementation
|
10 (16.4%)
|
7 (9.6%)
|
0.26
|
|
Reoperation
|
7 (11.5%)
|
8 (11.0%)
|
0.90
|
|
Readmission
|
9 (14.8%)
|
9 (12.3%)
|
0.65
|
|
Gross total resection
|
43 (84.3%)
|
37 (50.7%)
|
0.02[a]
|
|
At least one follow-up
|
59 (96.7%)
|
70 (95.9%)
|
0.99
|
Abbreviations: CSF, cerebrospinal fluid; DDAVP, desmopressin.
Note: Continuous variables reported as mean with standard error.
a Statistical significance.
Discussion
In this study, we characterize the socioeconomic and clinicopathologic differences
among the patient populations of a private, academic medical center and a prominent
public hospital affiliated with our institution. We identified several differences
in the baseline characteristics of these patient populations as well as differences
in management and postoperative outcomes. Our study is distinguished in its analysis
of patients from a single institution, whereas other efforts to uncover disparities
in transsphenoidal surgery have relied on large databases.[19]
[20]
[21] Furthermore, our study is unique in its inclusion of patients from two separate
hospital systems that share the same resident and attending physicians, which partially
mitigates differences in the operative experience of the two facilities. This is true
of all specialties involved in the management of pituitary tumors including neurosurgery,
otolaryngology, endocrinology, and ophthalmology. Previously, hospital-level surgical
volume has been identified as a driver of poorer outcomes after transsphenoidal surgery,
particularly for the development of diabetes insipidus and other electrolyte abnormalities.[21] Our results show that these disparities persist even when controlling for physicians'
operative experience, suggesting that these disparities are not fully explained by
volume alone.
We found that on average, the time-to-surgery from the initial neurosurgical recommendation
was 11.4 days lower for patients at our private hospital, though this difference was
not statistically significant. There are many factors that can contribute to wait
times. These include issues on the provider end of the spectrum such as operating
room availability and attending availability. They also include patient factors such
as navigating issues with time off work and managing family responsibilities. There
is also the complex interplay of patient and provider factors. For example, our institution's
attendings and trainees are typically aware that patients at our public hospital are
more likely to be lost to follow-up due to increased barriers to care, including lower
socioeconomic status, differences in education level, and language barriers. Physicians
may therefore feel more comfortable in recommending an observational approach with
serial imaging for private hospital patients, which would have the counterintuitive
effect of delaying surgery relative to patients at our public hospital. Interestingly,
more patients treated at our community hospital required emergency surgery. This is
consistent with previous findings that patients with decreased access to care are
more likely to present with pituitary apoplexy, an indication for emergency surgery.[24] Taken together, our results suggest that patients with impaired access to care may
present with more advanced disease; however, once care is established, there are minimal
differences in time-to-surgery.
The surgical management of patients at both hospitals was mostly similar; however,
an important procedural difference between our two facilities is the availability
of intraoperative magnetic resonance imaging (iMRI) at our private hospital. At NYU
Langone Health, iMRI is commonly performed to evaluate for residual tumor following
an initial attempt at resection, and in our cohort, iMRI was used in 60% of cases
at our private hospital. We observed no difference in operative time between our hospitals;
however, this may be partially explained by the inclusion of iMRI at NYU Langone Health
during surgery. Both the time used to obtain the image and any subsequent attempts
to resect additional tumor are included in operative time and may have masked a true
difference in operative time between the two hospitals. Furthermore, several previous
studies have documented higher rates of gross total resection when implementing this
technique[25]
[26]
[27]; therefore, iMRI could be partially responsible for the higher rate of gross total
resection achieved at our private hospital. Accordingly, this observed difference
in gross total resection may be the result of disparities in the technology and resources
available at our two hospitals, rather than the direct effect of socioeconomic differences
between their patient populations.
This study has several limitations, many of which are the result of its retrospective
design. As the data analyzed in this study were derived from medical records, it is
possible that some information was not documented. Given the shared faculty and house
staff, this is likely to have been comparable for both facilities, though we cannot
be certain whether this introduced bias into our data. Given the retrospective nature
of the study, the identification of postsurgical outcomes such as diabetes insipidus
was based on provider notes, which is likely to have introduced bias. Physicians at
both hospitals are encouraged to follow diagnostic algorithms for diabetes insipidus,
such as that described by Christ-Crain et al[28]; however, we cannot be certain as to whether these guidelines were strictly followed.
While our study assessed whether gross total resection was achieved on postoperative
imaging, other long-term outcomes such as recurrence were outside of the scope of
this study. Given the high density of hospitals in New York City, patients may also
have presented with delayed complications to other hospitals, which would not have
been detected during our data collection. Lastly, while our study was successful in
identifying socioeconomic differences between our hospitals' patient populations,
our sample is underpowered for multivariable analysis. As previously stated, several
studies using national databases have accomplished this.
There are myriad differences between our private and public hospitals that cannot
be controlled. At the public hospital, house staff often play a greater role in managing
patients both in and out of the operating room. Faculty have a higher volume of pituitary
surgery at the private hospital system and dedicated operative time for these cases,
whereas cases are scheduled on an as-needed basis at the public hospital. Private
hospitals have higher quality equipment and greater variety of modern surgical instruments
than public hospitals. Furthermore, the two hospital systems have different anesthesiologists
and operating room staff. These are just a few of the most readily apparent differences
between these institutions.
The aim of this study was not to determine whether presenting to a private hospital
or public hospital results in drastically different surgical outcomes; rather, our
goal was to characterize the patient populations and outcomes of those undergoing
transsphenoidal pituitary surgery of two socioeconomically disparate hospitals affiliated
with the same institution. The hospital to which a patient presents is likely a composite
of many socioeconomic factors including income, insurance status, U.S. citizenship,
and area of residence, all of which could contribute to health care disparities. Our
findings may inform practitioners by improving their awareness of these disparities,
allowing them to identify patients who are at the greatest risk for complications.
Conclusion
Socioeconomic disparities exist in the management and outcomes of pituitary tumors,
even when attempting to control for hospital faculty and staff. Compared with our
private hospital, patients treated at our public hospital were more likely to present
with visual field deficits and experience transient diabetes insipidus after undergoing
transsphenoidal surgery, and they were less likely to have gross total resection on
postoperative imaging. Further studies are needed to identify disparities in the long-term
outcomes of transsphenoidal surgery for pituitary tumors.
