Abstract
Introduction Higher surgical volumes correlate with superior patient outcomes for various surgical pathologies, including transnasal transsphenoidal (TNTS) pituitary tumor resection. With the introduction of endoscopic approaches, there have been nationwide shifts in technique with relative declines in microsurgery. We examined the volume-outcome relationship (VOR) for TNTS pituitary tumor surgery in an era of increasingly prevalent endoscopic approaches.
Methods Patients who underwent TNTS pituitary tumor resection between 2009 and 2011 were retrospectively identified in the State Inpatient Database subset of the Healthcare Cost and Utilization Project. Generalized linear mixed-effect models were used to assess odds of various outcome measures. Institutions were grouped into quartiles by case volume for analysis.
Results A total of 6,727 patients underwent TNTS pituitary tumor resection between 2009 and 2011. White or Asian American patients and those with private insurance were more likely to receive care at higher volume centers (HVC). Patients treated at HVC (>60 cases/year) were less likely to have nonroutine discharges (3.9 vs. 1.9%; p = 0.002) and had shorter length of stay (LOS; 4 vs. 2 days; p = 0.001). Overall, care at HVC trended toward lower rates of postoperative complications, for example, a 10-case/year increase correlated with a 10% decrease in the rate of iatrogenic panhypopituitarism (odds ratio [OR] = 0.90, 95% confidence interval [CI]: 0.81–0.99; p = 0.04) and 5% decrease in likelihood of diabetes insipidus (OR = 0.95, 95% CI: 0.90–0.99; p = 0.04) on multivariable analysis.
Conclusions Our analysis shows that increased case volume is related to superior perioperative outcomes for TNTS pituitary tumor resections. Despite the recent adoption of newer endoscopic techniques and concerns of technical learning curves, this VOR remains undisturbed.
Keywords: pituitary tumor, transsphenoidal, volume-outcome relationship
Introduction
There is abundant evidence for the correlation between high-volume centers or high-volume physician providers with superior outcomes for multiple surgical procedures in general surgery and oncology. 1 2 3 4 There has also been increasing evidence that patient mortality and morbidity are lower at higher volume centers when evaluating all neurosurgical cases, 5 as well as in the care of specific neurosurgical diseases, such as deep brain stimulation, 6 oncology, 7 8 and cerebrovascular pathology. 9 10 These results may be attributable to individual provider experience, lower risk patients, a higher proportion of elective admissions, and a more robust multidisciplinary care team at higher volume centers. 5 8 Very little data have been evaluated, in the English literature, with regard to the volume-outcome relationship (VOR) of transnasal transsphenoidal (TNTS) pituitary tumor resections. This may be, in part, due to investigators focusing more on comparing microscopic versus endoscopic transnasal approaches and outcomes.
The TNTS approach to masses in the sellar region was first introduced by Harvey Cushing in 1907, 11 but was largely abandoned because of poor visualization due to a narrow surgical window and poor-light penetration. The introduction of the microscope in 1962, by Jules Hardy, facilitated resurgence in the use of this technique with enhanced magnification and improved illumination. With the advent of endoscopic technology in the early 1990s, there began a significant shift in surgical approaches. The endoscope allows for a wider field of visualization, more instrument mobility, and angled views. 12 Comparisons of both microscopic and endoscopic techniques have shown that patients recover more quickly from an endoscopic approach and that it is associated with less immediate postoperative complications. 12 Concerns of difficulties with safety and efficacy of the endoscopic approach, especially with the learning curve associated with the transition from microsurgery, have been evaluated and studies have shown that extent of resection was similar in both, 13 14 operative times were shorter with endoscopic approaches, 14 and studies have either cited similar 14 or lower 12 13 rates of major complications with the endoscopic approach. One study showed that the learning curve required approximately 15 to 40 cases before a new surgeon became as facile with the endoscopic approach, having previously been experienced in the use of a microscope, 14 but despite this, the surgeries were just as safe. 12
Previous investigations supporting surgical experience as a major contributor to better outcomes for TNTS pituitary tumor resections have been published in single provider or single-center reports. 15 Barker et al has been the only group to investigate the VOR of TNTS pituitary tumor resections on a larger scale and they found that higher volume centers were associated with lower rates of short-term morbidity and mortality, especially for older patients. 16 This study evaluated data from the National Inpatient Sample (NIS) from 1996 to 2000, a time period during which use of the microscope still predominated in the United States. Since then, there has been a steady increase in the use of the endoscope in TNTS resections of pituitary pathology. 17 We sought to evaluate the VOR of TNTS pituitary tumor resections and short-term outcomes in the United States during a time period of increased use of, relatively newer, endoscopic techniques.
Methods
Data Source
A retrospective population-based cohort was obtained from the State Inpatient Database (SID) subset of the Healthcare Cost and Utilization Project (HCUP), Agency for Healthcare Research Quality. 18 This cohort of patients was obtained from the SID of Florida, Iowa, Washington, California, and New York for 2009 to 2011. The HCUP SID is an administrative dataset of deidentified patient-level data from hospital discharge records which includes over 100 variables and is reported across payer types. 18 Patients were included in the analysis if they were greater than 18 years of age and had an associated The International Classification of Diseases, Ninth revision, Clinical Modification (ICD-9CM) diagnosis code 227.3 (pituitary adenoma) and procedure code 07.14, 07.62 or 07.65 (transsphenoidal biopsy, partial excision, or total excision, respectively). Per HCUP data restriction, tabulation of queries with less than 10 events was omitted or presented in appropriately sized composite results to avoid patient identification in situations of rare events. The Institutional Review Board at our institution deemed the study exempt from review as the data are deidentified, protected, and publicly available.
Patient and Hospital Characteristics
Baseline demographic data were extracted from the SID, and included sex, race, insurance status, admission source (emergency room, other hospital, other health facility, long-term facility, court/law enforcement, and routine), and admission type (emergency, urgent, elective, and trauma center). Hospital characteristics including region, bed-size, and teaching status were assessed using the American Hospital Association (AHA) Annual Survey Database from 2011. 19 This nationwide database contains information categorizing an institution's organizational structure, facility and service lines, operation expenses, and staffing. Patient level data from HCUP SID was linked to the AHA annual surgery for hospital data. Annual hospital volume was first determined using the SID data. Hospital volume was then stratified by quartiles into low (1–8 cases/year), low–medium (9–25 cases/year), high–medium (26–60 cases/year), and high (>60 cases/year) volume centers.
Outcome Variables
The primary outcomes of this study were short-term postoperative complications, also identified using ICD-9CM codes. These include iatrogenic panhypopituitarism (IHYP, 253.7), diabetes insipidus (DI, 253.50), electrolyte abnormalities (276.00–276.52, 276.61–276.90276.00–276.52, and 276.61–276.90), infection (bacterial meningitis [320.0, 320.1, 320.2, 320.3, 320.7, 320.8, and 320.9] and intracranial abscess [324.0]), cerebrospinal fluid (CSF) rhinorrhea (394.81), neurological complications (997.00, 997.01, 997.02, and 997.09), hemorrhage/hematoma (intracerebral [430.00, 431.00, 432.0, 432.1, and 432.9], complicating a procedure [998.11 and 998.12]), cranial nerve deficits (paralytic ptosis [374.31] third/fourth/sixth palsy [378.50, 378.51, 378.52, 378.53, 378.54, 378.55, and 378.56], diplopia [368.20], ptosis of eyelid [374.30]), bitemporal hemianopia (368.47), cerebral arteriogram (88.41), mechanical ventilation (96.70, 96.71, and 96.72), blood transfusion (99.04), and deep venous thrombosis/pulmonary embolism (DVT [453.40, 453.41, and 453.42] or PE [415.00, 415.1, 415.11, 415.12, and 415.19] and inferior vena cava filter installation [38.7]).
Statistical Analysis
Patient and hospital characteristics were examined using descriptive statistics. To identify significant association of baseline independent variables with primary outcome, univariable generalized linear mixed-effects models were generated for each postoperative complication. Average annual case volume was incorporated into the model as a continuous variable, while patient age, sex, insurance status, teaching hospital status, and race were incorporated as categorical predictors. For postoperative complications panhypopituitarism and DI, multivariable models were also run adjusting for the above baseline variables. Random intercepts were allowed for each hospital facility in all models as described by the HCUP methods series and analysis guidelines. The composition of the final multivariable models was determined using the set of clinically relevant independent risk factors that best minimized Akaike's information criterion. An α error rate of p ≤ 0.05 was considered statistically significant and all analyses were conducted using SAS 9.4 (Cary, North Carolina, United States).
Results
A total of 6,727 patients who underwent TNTS surgery for resection of pituitary adenomas in the United States at 271 nonfederal hospitals between 2009 and 2011 were identified for analysis. Clinical characteristics of the patients are shown in Table 1 . The majority of patients were white (54%) and had private insurance (57%). The mean age was 52.3 years (range: 18–92 years) and there were slightly more females than males (52 vs. 48%). Sixty percent ( n = 4054) of cases were performed at teaching hospitals. Overall case volume remained stable over the 3 years ( Table 1 ). Median case volume across all centers was approximately 25 cases per year, with an interquartile range (IQR) of 8 to 60 cases/year. Case volumes were binned into quartiles, with most institutions in the “low” quartile (80%; Table 1 ).
Table 1. Clinical characteristics of study population and hospitals.
| Patient or case characteristics | Frequency (%), mean (SD) |
|---|---|
| Age | |
| Mean | 52.3 y (SD = 15.9) |
| Interquartile range | 39–67 y |
| Range | 18–92 y |
| Sex ( n = 6,466) | |
| Male | 3,073 (48) |
| Female | 3,393 (52) |
| Race/ethnicity ( n = 6,248) | |
| White | 3,373 (54) |
| Black | 879 (14) |
| Hispanic | 1,223 (20) |
| Asian or Pacific Islander | 425 (7) |
| Other | 348 (5) |
| Insurance ( n = 6,727) | |
| Medicare | 1,667 (25) |
| Medicaid | 766 (11) |
| Private | 3,805 (57) |
| Uninsured | 164 (2) |
| Other | 325 (5) |
| Teaching hospital ( n = 6,727) | |
| No | 2,673 (40) |
| Yes | 4,054 (60) |
| Case volume by year | |
| 2009 | 2,332 (34.7) |
| 2010 | 2,246 (33.4) |
| 2011 | 2,149 (31.9) |
| Hospital characteristics ( n = 271) | Frequency (%), median (IQR) |
| Institutional annual volume (cases/year) | 25 (8–60) |
| Low (1–8 cases/year) | 218 (80) |
| Low–medium (9–25 cases/year) | 33 (12) |
| High–medium (26–60 cases/year) | 15 (6) |
| High (>60 cases/year) | 5 (2) |
| Hospital beds per institution | 530 (346–815) |
Abbreviations: IQR, interquartile range; SD, standard deviation.
Outcomes, including length of stay (LOS), discharge disposition, mortality and short-term postoperative complications were also evaluated ( Table 2 ). The majority of patients were hospitalized for three days (IQR = 2–5). Most patients experienced routine discharges, home with no needs ( n = 6218; 92.4%) or with home health services ( n = 294; 4.4%). Overall mortality rate was 0.3% ( n = 22). The most prevalent complications included DI (11%, n = 720), neurological complications (8%, n = 505), cranial nerve deficits (5%, n = 336), and bitemporal hemianopsia (4%, n = 241; Table 2 ).
Table 2. Postoperative outcome measures.
| Outcomes | n (%) |
|---|---|
| Length of stay | |
| Median | 3 d |
| Interquartile range | 2–5 d |
| Discharge disposition ( n = 6705) | |
| Home/selfcare | 6,218 (92.4) |
| Home with home health service | 294 (4.4) |
| Transferred to short-term hospital | 23 (0.3) |
| Transferred to unidentified institution | 160 (2.4) |
| Left against medical advice | 10 (0.2) |
| Overall mortality | 22 (0.3) |
| Overall morbidity | 2,302 (34.2) |
| Endocrine complications | 835 (12.4) |
| Iatrogenic panhypopituitarism | 74 (1.1) |
| Diabetes insipidus | 720 (10.7) |
| Electrolyte abnormalities | 41 (0.6) |
| Neurological complications | 1,082 (16.1) |
| Cranial nerve deficits | 336 (5.0) |
| Bitemporal hemianopsia | 241 (3.6) |
| NOS | 505 (7.5) |
| Other postoperative complications (hemorrhage/hematoma, infection, CSF rhinorrhea, mechanical ventilation, and DVT/PE) | 385 (5.7) |
Abbreviations: CSF, cerebrospinal fluid; DVT, deep vein thrombosis; NOS, not otherwise specified; PE, pulmonary embolism.
Patient Characteristics and Outcomes at Low- versus High-Volume Centers
Table 3 shows clinical characteristics and outcomes of patients stratified by hospital center case-volume. There was no significant relationship between age and location of care received. However, Black and Hispanic patients are more likely to receive care at low-volume centers ( p < 0.001, Table 3 ). In addition, patients with “other” or no insurance are more likely to seek care at low volume centers, while patients with private insurance tended to receive care at higher volume centers.
Table 3. Patient characteristics and outcomes stratified by case volume.
| Case volume quartile | Low n = 1,719 (%) |
Low–medium n = 1,586 (%) |
High–medium n = 1,910 (%) |
High n = 1.512 (%) |
p -Value |
|---|---|---|---|---|---|
| Mean patient age in years a | 54 (SD =16.2) | 53 (SD = 15.5) | 52 (SD = 16.0) | 51 (SD = 15.8) | <0.001 |
| Insurance a | |||||
| Medicare (>65 years old) | 475/1,719 (27.6) | 383/1,586 (24.2) | 504/1,910 (26.4) | 305/1,512 (18.3) | <0.001 |
| Medicaid | 257/1,719 (15.0) | 202/1,586 (12.7) | 156/1,910 (8.2) | 1,51/1,512 (10.0) | |
| Private | 792/1,719 (46.1) | 879/1,586 (55.4) | 1,118/1,910 (58.5) | 1,016/1,512 (67.2) | |
| Other/uninsured | 195/1,719 (11.3) | 122/1,586 (7.7) | 132/1,910 (6.9) | 40/1,512 (2.6) | |
| Race a | |||||
| Black | 303/1,595 (19.0) | 223/1,490 (15.0) | 234/1,829 (12.8) | 119/1,334 (8.9) | <0.001 |
| Hispanic | 361/1,595 (22.6) | 329/1,490 (22.1) | 340/1,829 (18.6) | 193/1,334 (14.5) | |
| Asian or Pacific Islander | 110/1,595 (6.9) | 75/1,490 (5.0) | 135/1,829 (7.4) | 105/1334 (7.9%) | |
| Other (native American, other, unknown) | 87/1,595 (5.4) | 89/1,490 (6.0) | 91/1,829 (5.0) | 81/1,334 (6.1) | |
| White | 734/1,595 (46.0) | 771/1,490 (52.0) | 1,029/1,829 (56.3) | 836/1,334 (62.7) | |
| Median length of stay a (IQR) | 4 (3–7) | 3 (2–5) | 3 (2–4) | 2 (2–3) | <0.001 |
| Discharge disposition a | |||||
| Home ± home health services | 1,642/1,719 (95.5) | 1,531/1,586 (96.5) | 1,863/1,910 (97.5) | 1,476/1,512 (97.6) | 0.002 |
| Another institution (short-term hospital or unidentified) | 67/1,719 (3.9) | 46/1,586 (2.9) | 41/1,910 (2.1) | 29/1,512 (1.9) | |
| Post-operative complications a | |||||
| Endocrine a (Iatrogenic panhypopituitarism, diabetes insipidis, electrolyte abnormalities) | 269/1,719 (15.6) | 170/1,586 (10.7) | 261/1,910 (13.7) | 135/1,512 (8.9) | <0.001 |
| Neurologic a (NOS, cranial nerve deficits, bitemporal hemianopsia) | 279/1,719 (16.2) | 273/1,586 (17.2) | 343/1,910 (18.0) | 187 (12.3) | <0.001 |
| Other a (Infection, mechanical ventilation, DVT/PE, hemorrhage/hematoma, CSF rhinorrhea) | 98/1,719 (5.7) | 102/1,586 (6.4) | 120/1,910 (6.3) | 65/1,512 (4.3) | 0.02 |
Abbreviations: CSF, cerebrospinal fluid; DVT, deep vein thrombosis; IQR, interquartile range; NOS, not otherwise specified; PE, pulmonary embolism.
Statistically significant.
Patients treated at lower volume centers were more likely to require a nonroutine discharge ( p = 0.002, Table 3 ) and have longer LOS ( p = 0.001) compared with those treated at higher volume centers. There was also a trend toward lower rates of postoperative complications in higher volume centers, although none of these results remained significant on multivariable analysis. For example, on univariable analysis, patients at high-volume centers were 73% (odds ratio [OR]: 0.27, 95% confidence interval [CI]: 0.09–0.81; p = 0.02) less likely to have IHYP, 52% (OR = 0.48, 95% CI: 0.26–0.89; p = 0.02) less likely to have DI, and 77% (OR = 0.23, 95% CI: 0.09–0.64; p = 0.01) less likely to have electrolyte abnormalities compared with patients at low-volume centers.
Not specifically related to case-volume, but interesting to note, was the finding that patients at teaching hospitals were 43% (OR = 1.43, 95% CI: 1.02–2.01) more likely to have postoperative neurological complications coded for in the SID than those patients at other facility types ( p = 0.04). This was the only significant effect found on analysis of hospital teaching status and TNTS short-term complications.
Case-Volume Increases and Outcomes
To further investigate this relationship between case volume and perioperative outcomes (listed in Table 2 ), both univariable and multivariable analyses were performed to evaluate the effects of a 10-case increase in annual volume on postoperative complications. The most significant results were found when investigating IHYP ( Table 4 ) and DI ( Table 5 ). On univariable case-volume analysis, for every 10 additional cases, patients were 11% (OR = 0.89, 95% CI: 0.8–0.98; p = 0.02) less likely to have IHYP and 6% (OR = 0.94, 95% CI: 0.89–0.99; p = 0.03) less likely to have DI. Protective case-volume effects remained significant on multivariable analysis for IHYP, with every 10 additional cases conferring a 10% decrease in likelihood that patients would have IHYP (OR = 0.90, 95% CI: 0.81–0.99; p = 0.04) and 5% less likely to have DI (OR = 0.95, 95% CI: 0.90–0.99; p = 0.04). Univariable examination of other hospital-related complications, such as mechanical ventilation, DVT, or PE, did not reveal any significant case-volume effects.
Table 4. Univariable and multivariable generalized linear mixed effects model results for iatrogenic panhypopituitarism.
| Odds ratio (95% CI) | p -Value | Adjusted odds ratio (95% CI) | p -Value | |
|---|---|---|---|---|
| Age | 1.01 (0.99–1.02) | 0.28 | 0.99 (0.97–1.01) | 0.15 |
| Mean case volume (units = 10) | 0.89 (0.80–0.98) | 0.02 | 0.90 (0.81–0.99) | 0.04 |
| Sex (ref. = male) | 0.62 (0.39–0.99) | 0.05 | 0.62 (0.39–1.00) | 0.05 |
| Insurance | 0.02 | 0.04 | ||
| Medicare | 2.27 (1.35–3.82) | 0.002 | 2.92 (1.46–5.82) | 0.002 |
| Medicaid | 1.54 (0.71–3.33) | 0.28 | 1.44 (0.67–3.13) | 0.35 |
| Private (Ref) | – | – | ||
| Uninsured | 2.64 (0.89–7.85) | 0.08 | 1.74 (0.59–5.14) | 0.32 |
| Other | 0.69 (0.16–2.96) | 0.62 | 0.80 (0.19–3.45) | 0.77 |
| Race/ethnicity | 0.13 | |||
| White (ref.) | – | – | ||
| Black | 1.68 (0.92–3.05) | – | ||
| Hispanic | 0.57 (0.26–1.26) | – | ||
| Asian or Pacific Islander | 1.45 (0.59–3.54) | – | ||
| Other | 0.78 (0.24–2.59) | – | ||
| Teaching hospital | 0.74 (0.40–1.37) | 0.33 | – |
Abbreviations: CI, confidence interval; ref., reference.
Note: 1.1% ( n = 74) of the 6,466 patients in the multivariable model sample had iatrogenic panhypopituitarism.
Table 5. Univariable and multivariable generalized linear mixed effects model results for diabetes insipidus.
| Odds ratio (95% CI) | p -Value | Adjusted odds ratio (95% CI) | p -Value | |
|---|---|---|---|---|
| Age | 0.98 (0.97–0.98) | <0.001 | 0.97 (0.97–0.98) | <0.001 |
| Mean case volume (units = 10) | 0.94 (0.89–0.99) | 0.03 | 0.95 (0.90–0.99) | 0.04 |
| Sex (ref. = male) | 1.51 (1.29–1.78) | <0.001 | 1.33 (1.12–1.57) | <0.001 |
| Insurance | <0.001 | 0.10 | ||
| Medicare | 0.66 (0.53–0.81) | <0.001 | 1.20 (0.92–1.56) | 0.18 |
| Medicaid | 1.41 (1.10–1.80) | 0.01 | 1.31 (1.01–1.69) | 0.04 |
| Private (ref.) | – | – | ||
| Uninsured | 1.43 (0.90–2.27) | 0.13 | 1.38 (0.91–2.10) | 0.13 |
| Other | 1.35 (0.95–1.92) | 0.10 | 1.42 (0.96–2.12) | 0.08 |
| Race/ethnicity | 0.01 | 0.01 | ||
| White (ref.) | – | – | ||
| Black | 1.51 (1.19–1.90) | <0.001 | 1.38 (1.09–1.75) | 0.01 |
| Hispanic | 1.04 (0.83–1.31) | 0.72 | 0.85 (0.67–1.08) | 0.18 |
| Asian or Pacific Islander | 1.00 (0.71–1.41) | 0.99 | 0.99 (0.70–1.41) | 0.97 |
| Other | 1.26 (0.89–1.78) | 0.19 | 1.13 (0.79–1.61) | 0.50 |
| Teaching hospital | 0.93 (0.70–1.23) | 0.59 | – |
Abbreviations: CI, confidence interval; ref., reference.
Note: 11.1% ( n = 696) of the 6,248 patients in the multivariable model sample had diabetes insipidus.
Patient Characteristics and Outcomes
Certain patient demographics or characteristics were found to have a seemingly independent effect on perioperative outcomes. For example, patient age was found to have a significant influence on likelihood of nonroutine discharge. On both univariable and multivariable analyses, every one year increase in patient age resulted in an increased likelihood of nonroutine discharge (OR = 1.05, 95% CI 1.04–1.06, p < 0.001; adjusted odds ratio [AOR] = 1.03, 95% CI: 1.02–1.05, p < 0.001). Controlling for age, case volume, insurance type, and race, female patients are still 33% more likely to have DI than male patients (AOR = 1.33, 95% CI: 1.12–1.57; p = 0.04) but 38% less likely to have IHYP than male patients (AOR = 0.62, 95% CI: 0.39–1.00; p = 0.05). Black patients were also found to be 38% more likely to have DI than white patients (AOR = 1.38, 95% CI: 1.09–1.75; p = 0.01), even after controlling for case volume, sex, age, and insurance type.
After adjusting for case-volume, age, and sex, patients with nonprivate insurance were found to be more likely to suffer certain postoperative complications. For example, Medicare patients were 2.92 (95% CI: 1.46–5.82; p = 0.002) times more likely to have IHYP than those patients with private insurance, after controlling for case-volume, age, sex, and race. Medicare patients were also 53% (AOR = 1.53, 95% CI: 1.18–1.98; p = 0.001) more likely to have cranial nerve deficits, and 61% (AOR = 1.61, 95% CI: 1.19–2.18; p = 0.002) more likely to have bitemporal hemianopsia. Medicaid patients are also 31% (AOR = 1.31, 95% CI: 1.01–1.69; p = 0.04) more likely to have DI than patients with private insurance, on multivariable analysis. Both Medicare and Medicaid patients were more likely to experience nonroutine discharges than patients with private insurance. This relationship remained significant on multivariable analysis, Medicare (AOR = 3.27, 95% CI: 1.99–5.36, p < 0.001) and Medicaid (AOR = 5.49, 95% CI: 3.35–9.00, p < 0.001).
Discussion
The primary goal of this study was to reexamine the VOR of TNTS pituitary tumor resections in an era of nationwide shifts in surgical techniques. Prior investigation into the VOR of TNTS pituitary resections took place during a time (1996–2000) in which the use of endoscopy had only recently been introduced. 16 A nationwide evaluation of Current Procedural Terminology (CPT) codes showed a significant and steady increase in endoscopic procedures (130.2 cases/year) versus decrease in microscopic (84.6 cases/year) procedures from 2003 to 2013. 17 Therefore, our analyses of the SID databases for TNTS pituitary adenoma resection in the United States, between 2009 and 2011, examines the VOR amidst increasingly prevalent endoscopic approaches and relative declines in microsurgery.
We studied 6,727 patients who underwent TNTS pituitary tumor resections at 271 U.S. institutions in our analysis and found that patients of higher volume hospitals had lower rates of complications compared with those treated at lower volume centers. In fact, a 10-case/year increase was associated with a reduction in the rate of postoperative complications, in particular a 10% reduction in IHYP and a 5% reduction in DI. Patients at these higher volume centers were also more likely to be insured by private insurance, identify as white, and have a shorter LOS. Although our analysis did not directly assess the costs of hospitalization, decreased adverse short-term outcomes, decreased LOS, and increased routine, discharges associated with higher-volume hospitals would likely impact overall costs of care. Future investigations may seek to highlight any associations institutional volume may have on hospitalization costs.
Although there was no significant effect of hospital volume on overall mortality, it is interesting to note that overall mortality during our study period (2009–2011) was 0.3% ( n = 22/6727) versus 0.6% as cited by Barker et al over the period of 1996 to 2000 ( n = 32/5497). 16 This suggests that despite increasing adoption of newer endoscopic techniques, there has been no increase in mortality. Rates of discharge to home were comparable between our two studies as well, with 96.8% ( n = 6,152/6,727) of our patients discharged home with or without home health services compared with 96.3% cited by Barker et al. 16 Our findings also corroborated the trend of more routine discharges at higher volume centers. Our patients were of comparable age (mean 52.3 ± 15.9 years vs. 50 years) and of similar distribution of insurance, gender, and race. 16 Similar to Barker et al, we found that age and public insurance were significant predictors of worse outcome at hospital discharge.
Interestingly, our results indicated that patients receiving care at teaching hospitals were more likely to have postoperative neurological complications. This may be due to increased identification and documentation of even minor complications due to the emphasis on resident education at these institutions. This may also be due to a larger proportion of more complex cases or larger tumors being treated at academic institutions. When examining patient characteristics, female gender, non-White race, and public insurance were also associated with increased rates of certain postoperative complications, independent of volume of case. Although, this was not the focus of our paper, these results highlight some potential population-based risk factors that may influence the likelihood of poor short-term outcomes after TNTS pituitary tumor resection.
Investigation of administrative databases is associated with some obvious limitations of surveillance bias, data limitation to inpatient encounters only, and inherent limitations due to reliance on billing codes that may be inaccurate or incomplete. 20 21 In addition, restricted tabulation of queries with less than 10 events, per HCUP policy, may have led to desensitization of certain findings. Despite this, administrative data plays a valuable role in health care research and serves as a robust source of information on rare conditions. Other limitations of our study include the lack of assessment of tumor size, and pre- versus postoperative vision status, as these likely play a role in influencing short-term postoperative outcomes and surgical efficacy. Unfortunately, coding that would differentiate endoscopic versus microscopic approaches were not available for differentiation and thus, the exact prevalence of the use of endoscopic techniques is unknown but can be inferred from the analysis by Rolston et al to have increased proportionately over this time period. 17 Future database investigations may also seek to analyze patient outcomes stratified by surgical approach.
Future studies of the VOR of TNTS pituitary tumor resections may also consider stratification of outcomes by individual provider volume. Although this information was not available on our analysis, prior studies have shown that individual provider/surgeon volume is responsible for a significant proportion of the institutional VOR. 3 22
Conclusion
We conducted an investigation of a large representative cohort of patients in the United States who underwent TNTS pituitary tumor resection during 2009 to 2011. Our analysis revealed a significant relationship between higher institutional case volume and lower frequencies of postoperative complications, nonroutine discharges, and shorter LOS. Although we were unable to investigate the VOR in terms of its impact on surgical efficacy (i.e., extent of tumor resection and visual outcomes), or stratify these results by surgical approach, our results still have significance for patients with pituitary tumors and the providers or insurers that are involved in their care. Overall, despite changes in the overall landscape of TNTS pituitary tumor resection, such as increases in the prevalence of relatively newer endoscopic techniques, the positive impact of care received at higher volume centers on perioperative outcomes remains unperturbed.
Footnotes
Conflict of Interest None declared.
References
- 1.Cobb A N, Wong Y M, Brownlee S A. Perioperative support, not volume, is necessary to optimize outcomes in surgical management of necrotizing enterocolitis. Am J Surg. 2017;213(03):502–506. doi: 10.1016/j.amjsurg.2016.11.014. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Bach P B, Cramer L D, Schrag D, Downey R J, Gelfand S E, Begg C B. The influence of hospital volume on survival after resection for lung cancer. N Engl J Med. 2001;345(03):181–188. doi: 10.1056/NEJM200107193450306. [DOI] [PubMed] [Google Scholar]
- 3.Birkmeyer J D, Stukel T A, Siewers A E, Goodney P P, Wennberg D E, Lucas F L. Surgeon volume and operative mortality in the United States. N Engl J Med. 2003;349(22):2117–2127. doi: 10.1056/NEJMsa035205. [DOI] [PubMed] [Google Scholar]
- 4.Dimick J B, Upchurch G R., Jr Endovascular technology, hospital volume, and mortality with abdominal aortic aneurysm surgery. J Vasc Surg. 2008;47(06):1150–1154. doi: 10.1016/j.jvs.2008.01.054. [DOI] [PubMed] [Google Scholar]
- 5.Davies J M, Ozpinar A, Lawton M T. Volume-outcome relationships in neurosurgery. Neurosurg Clin N Am. 2015;26(02):207–218. doi: 10.1016/j.nec.2014.11.015. [DOI] [PubMed] [Google Scholar]
- 6.Kalakoti P, Ahmed O, Bollam P, Missios S, Wilden J, Nanda A. Predictors of unfavorable outcomes following deep brain stimulation for movement disorders and the effect of hospital case volume on outcomes: an analysis of 33, 642 patients across 234 US hospitals using the National (Nationwide) Inpatient Sample from 2002 to 2011. Neurosurg Focus. 2015;38(06):E4. doi: 10.3171/2015.3.FOCUS1547. [DOI] [PubMed] [Google Scholar]
- 7.De la Garza-Ramos R, Abt N B, Kerezoudis P, Krauss W, Bydon M. Provider volume and short-term outcomes following surgery for spinal metastases. J Clin Neurosci. 2016;24:43–46. doi: 10.1016/j.jocn.2015.08.008. [DOI] [PubMed] [Google Scholar]
- 8.Trinh V T, Davies J M, Berger M S. Surgery for primary supratentorial brain tumors in the United States, 2000-2009: effect of provider and hospital caseload on complication rates. J Neurosurg. 2015;122(02):280–296. doi: 10.3171/2014.9.JNS131648. [DOI] [PubMed] [Google Scholar]
- 9.Rinaldo L, McCutcheon B A, Murphy M E. Quantitative analysis of the effect of institutional case volume on complications after surgical clipping of unruptured aneurysms. J Neurosurg. 2017;127(06):1297–1306. doi: 10.3171/2016.9.JNS161875. [DOI] [PubMed] [Google Scholar]
- 10.Bekelis K, Connolly I D, Do H M, Choudhri O. Operative volume and outcomes of cerebrovascular neurosurgery in children. J Neurosurg Pediatr. 2016;18(05):623–628. doi: 10.3171/2016.5.PEDS16137. [DOI] [PubMed] [Google Scholar]
- 11.O'Malley B W, Jr, Grady M S, Gabel B C. Comparison of endoscopic and microscopic removal of pituitary adenomas: single-surgeon experience and the learning curve. Neurosurg Focus. 2008;25(06):E10. doi: 10.3171/FOC.2008.25.12.E10. [DOI] [PubMed] [Google Scholar]
- 12.Singh H, Essayed W I, Cohen-Gadol A, Zada G, Schwartz T H. Resection of pituitary tumors: endoscopic versus microscopic. J Neurooncol. 2016;130(02):309–317. doi: 10.1007/s11060-016-2124-y. [DOI] [PubMed] [Google Scholar]
- 13.Zaidi H A, Awad A W, Bohl M A. Comparison of outcomes between a less experienced surgeon using a fully endoscopic technique and a very experienced surgeon using a microscopic transsphenoidal technique for pituitary adenoma. J Neurosurg. 2016;124(03):596–604. doi: 10.3171/2015.4.JNS15102. [DOI] [PubMed] [Google Scholar]
- 14.Eseonu C I, ReFaey K, Rincon-Torroella J. Endoscopic versus microscopic transsphenoidal approach for pituitary adenomas: Comparison of outcomes during the transition of methods of a single-surgeon. World Neurosurg. 2017;97:317–325. doi: 10.1016/j.wneu.2016.09.120. [DOI] [PubMed] [Google Scholar]
- 15.Shahlaie K, McLaughlin N, Kassam A B, Kelly D F. The role of outcomes data for assessing the expertise of a pituitary surgeon. Curr Opin Endocrinol Diabetes Obes. 2010;17(04):369–376. doi: 10.1097/MED.0b013e32833abcba. [DOI] [PubMed] [Google Scholar]
- 16.Barker F G, II, Klibanski A, Swearingen B. Transsphenoidal surgery for pituitary tumors in the United States, 1996-2000: mortality, morbidity, and the effects of hospital and surgeon volume. J Clin Endocrinol Metab. 2003;88(10):4709–4719. doi: 10.1210/jc.2003-030461. [DOI] [PubMed] [Google Scholar]
- 17.Rolston J D, Han S J, Aghi M K. Nationwide shift from microscopic to endoscopic transsphenoidal pituitary surgery. Pituitary. 2016;19(03):248–250. doi: 10.1007/s11102-015-0685-y. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Steiner C, Elixhauser A, Schnaier J. The healthcare cost and utilization project: an overview. Eff Clin Pract. 2002;5(03):143–151. [PubMed] [Google Scholar]
- 19.American Hospital Association The AHA Annual Survey DatabaseAccessed September 18, 2019 at:https://www.ahadataviewer.com/additional-data-products/AHA-Survey/
- 20.Haut E R, Pronovost P J, Schneider E B.Limitations of administrative databases JAMA 2012307242589–2590., author reply 2589–2590 [DOI] [PubMed] [Google Scholar]
- 21.Stulberg J J, Haut E R. Practical guide to surgical data sets: healthcare cost and utilization project national inpatient sample (NIS) JAMA Surg. 2018;153(06):586–587. doi: 10.1001/jamasurg.2018.0542. [DOI] [PubMed] [Google Scholar]
- 22.Ward B K, Gourin C G, Francis H W. Vestibular schwannoma surgical volume and short-term outcomes in Maryland. Arch Otolaryngol Head Neck Surg. 2012;138(06):577–583. doi: 10.1001/archoto.2012.877. [DOI] [PubMed] [Google Scholar]