| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Brain Tumor Center (F.G.B., B.S.), Neuroendocrine Clinical Center (A.K., B.S.), Neurosurgical Service, Departments of Surgery (F.G.B., B.S.) and Medicine (A.K.), Massachusetts General Hospital, and Harvard Medical School, Boston, Massachusetts 02114
Address all correspondence and requests for reprints to: Brooke Swearingen, M.D., ACC 331 Massachusetts General Hospital Fruit Street, Boston, Massachusetts 02114. E-mail: swearingen{at}helix.mgh.harvard.edu.
 |
Abstract |
|---|
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionConclusionsReferences |
|---|
A total of 5497 operations were performed at 538 hospitals by 825 surgeons. Outcome measured at hospital discharge was: death (0.6%), discharge to long-term care (0.9%), to short-term rehabilitation (2.1%), or directly home (96.2%). Outcomes were better after surgery at higher-volume hospitals (OR 0.74 for 5-fold-larger caseload, P = 0.007) or by higher-volume surgeons (OR 0.62, P = 0.02). A total of 5.4% of patients were not discharged directly home from lowest-volume-quartile hospitals, compared with 2.6% at highest-volume-quartile hospitals. In-hospital mortality was lower with higher-volume hospitals (P = 0.03) and surgeons (P = 0.09). Mortality rates were 0.9% at lowest-caseload-quartile hospitals and 0.4% at highest-volume-quartile hospitals. Postoperative complications (26.5% of admissions) were less frequent with higher-volume hospitals (P = 0.03) or surgeons (P = 0.005). Length of stay was shorter with high-volume hospitals (P = 0.02) and surgeons (P < 0.001). Hospital charges were lower for high-volume hospitals, but not significantly.
This analysis suggests that higher-volume hospitals and surgeons provide superior short-term outcomes after transsphenoidal pituitary tumor surgery with shorter lengths of stay and a trend toward lower charges.
|
Introduction |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionConclusionsReferences |
|---|
Surgical excision is the initial treatment for most pituitary tumors, excluding prolactinomas, for which surgery is often indicated when medical therapy is unsuccessful. Although reports of outcomes after transsphenoidal surgery for pituitary tumors usually originate in centers that specialize in such surgery, the extent to which transsphenoidal surgery is currently performed at low-volume centers and the results achieved in nonspecialized settings are relatively unknown.
We studied the volume-outcome relationship for transsphenoidal surgery for pituitary tumors performed in a representative sample of U.S. hospitals between 1996 and 2000. Specifically, we related the chance of adverse outcomes [in-hospital mortality, outcome at hospital discharge, complications of surgery, length of stay (LOS), and hospital charges] to the annual hospital and surgeon caseload of transsphenoidal surgery, as well as to other patient and provider characteristics.
|
Subjects and Methods |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionConclusionsReferences |
|---|
Inclusion and exclusion criteria and definition of endpoints
We defined an admission for transsphenoidal surgery for pituitary tumor by using a combination of ICD-9-CM diagnosis and primary procedure codes. We required a primary procedure code of 07.14, 07.62, or 07.65 (biopsy, partial resection, or total resection of pituitary gland using transsphenoidal approach, respectively) and one or more of the following diagnosis codes: 227.3 (benign neoplasm of pituitary), 194.3 (malignant neoplasm of pituitary), 237.0 (pituitary neoplasm of uncertain behavior), 239.7 (endocrine neoplasm of uncertain nature), 253.0 (acromegaly), or 255.0 (Cushing’s syndrome). Other intrasellar lesions, such as craniopharyngiomas and Rathke’s cleft cysts, were excluded.
Two primary endpoints were examined: in-hospital mortality and discharge to institutions other than home. In-hospital mortality was coded directly in the NIS database and was analyzed using logistic regression. Discharge to institutions other than home was coded on a four-level scale and was analyzed using ordinal logistic regression, which allows use of the entire spectrum of outcomes rather than simplifying to a single cutpoint with resultant information loss (8, 9, 10). Discharge was coded as death, discharge to a long-term facility, discharge to other facilities, or discharge home, as follows. NIS data distinguishes discharge to long-term facilities (such as skilled nursing facilities) from discharge to other (intermediate or short-term care) facilities for all states except California and Maryland; for these states, we coded these discharges (0.3% of the total) as discharge to other facilities. We counted discharge home with home health care or iv therapy (4.5% of discharges) and discharge against medical advice (0.05% of discharges) as discharge home. Discharge to another acute care hospital (0.2% of discharges) was counted as discharge to an institution other than home, not as discharge to a long-term facility.
LOS and total hospital charges were coded in NIS data. Four patients with LOS of zero were recoded as missing (all were discharged alive). LOS and hospital charge analyses included only patients discharged from hospital alive. LOS and hospital charge data were highly positively skewed and were analyzed as the logarithmic transforms.
Patient characteristics
Patient age, sex, race, median household income for ZIP code of residence (11), primary payer (Medicare, Medicaid, private insurance, self-pay, no charge, other), type of admission (emergency, urgent, elective) and admission source (emergency room, transfer from another hospital, transfer from long-term care, and routine) were coded in NIS data. Eleven patients (0.2%) with admission type listed as other were recoded as routine admissions. Five patients (0.1%) with admission source of court/law enforcement were recoded as admissions from home. More than 5% of discharges had missing values for three variables used principally as stratification factors for other analyses, race (22% missing), admission type (21% missing), and whether the principal procedure was performed on the first hospital day (12% missing). When these variables were used as stratification factors, missing values for race and admission type were imputed as follows. Missing race was set to white. Missing admission type was set to emergency for admissions whose source was the emergency room, to urgent for admissions that were transfers from another hospital, and to routine for admissions from other sources. Whether the principal procedure was performed on the first hospital day was not imputed, and when race or admission type were the focus of the analysis, imputed values were not used.
To assess the effect of general medical comorbidity, the set of 30 medical comorbidity markers described by Elixhauser et al. (12), excluding the two specific neurological comorbidity variables (paralysis and other neurological deficit) and three comorbidity variables likely to represent postoperative conditions (fluid and electrolyte disorders, blood loss anemia, and deficiency anemias), were calculated using Agency for Healthcare Research and Quality software (www. ahcpr.gov/data/hcup/comorbid.htm) and summed to give a single comorbidity score ranging between 0 and 25.
Endocrine diagnoses were defined as follows: Cushing’s disease (ICD-9-CM 255.0), acromegaly (253.0), and noniatrogenic panhypopituitarism (253.2). Visual loss likely to be due to pituitary tumors was defined as visual field defects (368.40–49), optic atrophy (377.10), compression of optic nerve(s) (377.49), disorders of the optic chiasm associated with pituitary neoplasm (377.51), or disorders of other visual pathways due to neoplasm (377.61). We identified potential complications of transsphenoidal surgery using the following codes: postoperative neurological complications, including those due to infarction or hemorrhage (997.00–997.09); hematoma complicating a procedure (998.1–998.13); any intracerebral hemorrhagic event (430–432 or 998.1–998.13); fluid and electrolyte abnormalities (276.0-.9); diabetes insipidus (DI, 253.5); iatrogenic panhypopituitarism (253.7); diplopia, ptosis, or deficits of cranial nerves 3, 4, or 6 (368.2, 374.30–31, 378.50–59); cerebrospinal fluid rhinorrhea (349.81); performance of a cerebral arteriogram (88.41); mechanical ventilation (96.70–96.72); deep venous thrombosis, pulmonary embolism, or placement of an inferior vena cava filter (415, 415.11–19, 453.8–9, 38.7), and transfusion of packed red blood cells (RBCs) (99.04).
Provider and hospital characteristics
Hospital region (Northeast, Midwest, South, or West), location (rural or urban), teaching status, and bed size (small, medium, large) were coded in NIS data. We derived hospital and surgeon volumes of transsphenoidal surgery by counting the cases for each identified surgeon and hospital in the database. Because hospital and physician volumes were positively skewed, the logarithmic transforms were used when volume measures were entered into regression models.
Statistical methods
Statistical methods included the Fisher’s exact and Wilcoxon rank tests, Spearman rank correlation, and loglinear least-squares, ordinary logistic, and proportional-odds ordinal logistic regression (13, 14, 15). To correct for possible clustering of similar outcomes within hospitals, which could cause falsely inflated estimates of the statistical significance of regression coefficients, a sandwich variance-covariance matrix was estimated from the data using methods due to Huber and White, with adjustment for clustering by hospital or surgeon (15). LOS and hospital charges were analyzed as logarithmic transforms using least-squares regression corrected for clustering as described above. Calculations were performed using SAS (version 8.2; SAS Institute, Cary, NC) and S-plus (Version 3.3 for Windows; Insightful, Inc., Seattle, WA) with the Hmisc and Design modeling function software libraries due to Harrell [15 ; Hmisc and Design libraries for S-Plus for Windows (2000)—software and electronic documentation available from http://hesweb1.med. virginia.edu/biostat/s/splus.html] and the LOCFIT local-likelihood regression library due to Loader (16, 17). Extrapolations to the entire U.S. population were adjusted for the NIS stratified survey method using SAS PROC SURVEYMEANS (18). All P values are two-tailed.
|
Results |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionConclusionsReferences |
|---|
. Most patients were white, age 35–65 yr, with slightly more women than men. Cushing’s disease was coded in 7% of patients and acromegaly in 6%. Most admissions were classified as routine and admission through the emergency ward or as a transfer from another hospital was uncommon. There was no upward or downward trend in the annual number of cases during the study period and no trend in patient age, mortality, or outcome at hospital discharge over time. Thirty-two patients died (0.6%), 52 (0.9%) were discharged to long-term care facilities such as skilled nursing facilities, 118 (2.1%) were discharged to other facilities such as rehabilitation hospitals, and 5290 patients were discharged home (96.3%).
|
We tested five demographic variables as potential outcome predictors: age, sex, race, income in the ZIP code of the patient’s residence (a surrogate for patient’s income), and primary payer for care. In univariate analyses, age was a significant predictor both of mortality [odds ratio (OR) 1.6 per decade, 95% confidence interval (CI) 1.2–2.1, P < 0.001] and of worse outcome at hospital discharge (OR 1.8, 95% CI 1.6–2.0, P < 0.001; Fig. 1). Patient gender and race (coded as white vs. nonwhite or black vs. nonblack) had no significant relationship with mortality or outcome. Median income in the patient’s ZIP code of residence was significantly related to outcome (better outcome in higher income areas, OR 0.86 95% CI 0.75–0.98, P = 0.02), but was not significantly related to mortality (P = 0.3). Primary payer for care was also significantly related both to mortality and to outcome at hospital discharge, findings that persisted after adjustment for patient age. Private insurance status predicted lower mortality (OR 0.32, 95% CI 0.12–0.80, P = 0.01) and better outcome at discharge (OR 0.19, 95% CI 0.09–0.44, P < 0.001). Age, sex, median income in ZIP code of residence, and stratification by race, primary payer, and geographic region were included in all models described below.
|
We tested coded diagnoses of Cushing’s disease (coded in 7% of the cohort), acromegaly (6%), noniatrogenic panhypopituitarism (6%), and visual loss (6%) as outcome predictors. Cushing’s disease was associated with trends toward higher mortality (OR 3.5, 95% CI 0.95–13, P = 0.06) and worse outcome at hospital discharge (OR 1.8, 95% CI 0.9–3.5, P = 0.1). Acromegaly, noniatrogenic panhypopituitarism, and visual loss were not significant predictors of mortality or outcome at discharge. Presence or absence of Cushing’s disease was included in all subsequent models.
Hospital and surgeon characteristics and outcome
Patients were treated at 538 hospitals. For 2727 patients (50% of the total), 825 treating surgeons were identified in the database. Between 95 and 98% of surgeons operated at one hospital, and 2–5% operated at two or three hospitals. (Because not all hospitals per state are sampled, this represents a minimum estimate.) Hospitals and surgeons varied widely in the volume of transsphenoidal pituitary tumor surgery reported. Analyzed on a per-patient basis, the median annual number of transsphenoidal pituitary tumor operations was 10 per hospital (range, 1–126 admissions; 25th percentile, five admissions; 75th percentile, 25 admissions) or three per surgeon (range, 1–33 admissions; 25th percentile, two admissions; 75th percentile, seven admissions). For 251 patients (5%), no other transsphenoidal pituitary tumor operation was reported during that year at their hospital, and for 627 patients (23%) no other transsphenoidal pituitary tumor operation was reported that year by their surgeon. Table 2 shows clinical characteristics of patients treated at hospitals in the lowest- and highest-volume quartiles for transsphenoidal pituitary tumor surgery caseload (one to four admissions per year and 25 or more admissions per year, respectively).
|
shows the probability of discharge other than to home for operations at hospitals grouped by caseload into four quartiles and for surgeons similarly grouped. Discharge other than to home occurred after 5.4% (71/1312) of admissions for transsphenoidal pituitary tumor surgery at hospitals with annual caseload of one to four admissions compared with 2.6% (36/1390) of admissions at hospitals with annual caseload of 25 admissions or more. The corresponding rates for surgeon caseload were 6.2% (39/627) for admissions to surgeons with annual caseload of one case and 2.7% (18/657) for admissions to surgeons with annual caseload of eight or more.
|
shows the probability of discharge other than to home as a function of age for highest- and lowest-volume quartile hospitals. Older patients had a much greater difference between highest- and lowest-volume-quartile hospitals in probability of discharge other than to home.
|
shows mortality by hospital and surgeon caseload quartile. 0.9% of patients operated at lowest-caseload-quartile hospitals died, compared with 0.4% at highest-caseload-quartile hospitals. A total of 1.1% of patients whose surgeons were in the lowest caseload quartile died, compared with 0.2% for patients of highest-caseload-quartile surgeons.
|
We tested patient characteristics as potential predictors of care at high-volume hospitals (Table 2
) or by high-volume surgeons. Older patients were less likely to have a high-volume hospital (P < 0.001 or surgeon (P = 0.003). Race was a significant predictor of hospital volume (P < 0.001) and surgeon volume (P = 0.03). Hospital and surgeon caseloads were higher for white patients than for black patients. Primary payer for care was a significant predictor of both hospital (P < 0.001) and surgeon volume (P = 0.04), with highest provider volumes for those who had private insurance. Patients from higher-income areas of residence had higher-volume hospitals (P < 0.001), but not surgeons (P = 0.3). Emergency or urgent admissions were more common to lower-caseload surgeons (P = 0.02) but not significantly related to hospital caseload (P = 0.3). Patients with more medical comorbidity tended to have lower-volume hospitals (P = 0.002) and surgeons (P = 0.06). Patients with Cushing’s disease were more commonly treated at high-volume hospitals and by high-caseload surgeons (P < 0.001 for both), as were patients with acromegaly (P = 0.08 for hospitals and P = 0.03 for surgeons). (Some of these admissions may have represented second operations at high-volume centers after unsuccessful surgery elsewhere, but this could not be determined from the database.) Patients with noniatrogenic panhypopituitarism were more commonly treated at low-volume hospitals (P < 0.001) and by low-volume surgeons (P = 0.008), as were patients with visual loss (P < 0.001 for hospitals and P = 0.001 for surgeons).
Complications and provider volume
We studied several complications of surgery or perioperative care: postoperative neurological complications, including those due to infarction or hemorrhage (reported in 254/5497 patients, 4.6%), hematoma complicating a procedure (86 patients, 1.6%), any intracranial hemorrhage (126 patients, 2.3%), mechanical ventilation (63 patients, 1.1%), fluid and electrolyte abnormalities (486 patients, 8.8%), DI (578 patients, 10.5%), iatrogenic panhypopituitarism (46 patients, 0.8%), performance of a cerebral arteriogram (70 patients, 1.3%), cerebrospinal fluid rhinorrhea (77 patients, 1.4%), cranial nerve 3, 4, or 6 palsies (176 patients, 3.2%) postoperative thrombotic disorders (deep venous thrombosis, pulmonary embolism, or placement of an inferior vena cava filter; 31 patients, 0.6%), and transfusion of packed RBCs (63 patients, 1.1%). We omitted noniatrogenic hypopituitarism, coded in 46 patients (0.8%), from these analyses because this condition is almost always identified after hospital discharge; this code was not significantly associated with mortality or discharge disposition (Table 3) or with hospital or surgeon caseload (data not shown), and repeat analyses including this code in the aggregate definition of complications gave essentially identical results.
|
).
In multivariate analysis, a complication was more likely in patients with more medical comorbidity (P < 0.001) or Cushing’s disease (P = 0.06). Larger provider caseloads were associated with less frequent complications. Odds ratios for occurrence of one or more complications were 0.77 for a 5-fold larger hospital caseload (95% CI 0.61- 0.97, P = 0.03) and 0.76 for a 5-fold larger surgeon caseload (95% CI 0.65–0.89, P = 0.005). Complications occurred during 31.1% of admissions to lowest-quartile-volume hospitals, compared with 23.0% of admissions to highest-quartile-volume hospitals; corresponding rates for surgeon volume were 33.2% (lowest volume) and 23.7% (highest volume; Fig. 5). After adjustment for hospital caseload, hospital teaching status and bed size were not significantly related to complication probability.
|
and 5). Higher hospital and surgeon caseload were both significantly associated with lower rates of fluid and electrolyte abnormalities (P < 0.004 for both) and DI (P < 0.03 for both), and there was a trend toward less frequent mechanical ventilation at higher-volume hospitals (P = 0.06).
|
|
LOS decreased significantly during the study period (by 4.6% per year, P < 0.001), although the median was 4 d both in 1996 and in 2000. After multivariate adjustment for the variables described above and stratification by treatment year, LOS was significantly shorter at larger-volume hospitals (P = 0.02). Adjusted for hospital caseload, neither hospital teaching status nor hospital bed size was correlated with LOS. In a similar multivariate model, larger surgeon caseload was also associated with shorter LOS (P < 0.001).
Total hospital charges increased significantly during the study period, from a median of $17,100 in 1996 to $20,200 in 2000 (by 4.3% per year, P = 0.02). After multivariate adjustment for the variables described above and stratification by treatment year, there was a trend toward lower charges at higher-volume hospitals (7.6% lower charges for a 5-fold larger caseload, 95% CI 5% higher to 18% lower, P = 0.2). Adjusted for hospital caseload, neither hospital teaching status nor hospital bed size was correlated with charges. The relationship between surgeon caseload and hospital charges was not significant (P = 0.7).
|
Discussion |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionConclusionsReferences |
|---|
Surgical excision is the primary treatment for most pituitary tumors. Our data indicate that adverse outcomes are less common after pituitary surgery by higher-volume providers. Most of the differences in outcome we found were statistically significant after casemix adjustment using patient-related risk factors such as age, comorbidity, and Cushing’s disease, previously reported to be associated with adverse outcomes (19, 20). High-volume provider care was also associated with significantly shorter hospital stays and a trend toward lower total hospital charges.
Better patient outcome after transsphenoidal pituitary tumor surgery by a more experienced surgeon was first reported by Cushing, who noted a progressive decrease in mortality rates from 40% (multiple earlier surgeons), to 13.7% in his own early series (21), to 3.9% by the end of his career (22). In modern series reported by specialist surgeons, transsphenoidal pituitary surgery mortality rates are about 1% or less (19, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32). These series may not reflect results in less specialized practice settings, and the extent to which pituitary surgery is concentrated in specialized centers and the results achieved outside these settings both remain largely unknown. We found that most transsphenoidal pituitary surgery in the United States during 1996–2000 was performed in low-volume hospitals, by low-volume surgeons. The median annual hospital caseload was 10 admissions, and the median annual surgeon caseload was three admissions. Almost one quarter of patients represented the only transsphenoidal pituitary operation in the database that year for their surgeon.
Recent series have suggested that surgeons who are pituitary tumor specialists have higher endocrine cure rates for Cushing’s disease and acromegaly (25, 27, 33). When one surgeon or a few operate at a specialized center, larger cumulative personal experiences result, with better outcomes—the surgical learning curve (34). Better acromegaly cure rates with greater experience were reported by one surgeon operating in Oxford, United Kingdom, 1974–1995 (35) and by three surgeons operating in Boston, Massachusetts, 1978–1996 (23). We found no evidence that having only one pituitary surgeon per institution improved outcome, other than through larger personal caseload.
Provider caseload effects on mortality or operative morbidity after transsphenoidal surgery are more difficult to detect in single-institution studies because these endpoints are infrequent. Rees et al. (25) reported lower hypopituitarism rates when Cushing’s disease surgery shifted from three surgeons to one, but there was no clear decrease in other complications. Ciric et al. (36) reported a survey study on transsphenoidal surgery complications, based on 958 responding U.S. surgeons (37% response rate). Surgeons with larger cumulative experience reported significantly lower rates of 13 of 14 complications assessed. We found a weaker correlation between complication rates and surgical caseload than did Ciric et al. (36). Possible explanations include undercoding or miscoding of complications in NIS data, reducing statistical power, or underreporting of complications by experienced surgeons, overreporting by inexperienced surgeons, or both in the survey study.
Our enforced use of outcomes measured during hospitalization means that our rates of some complications differ from those previously reported. The 26.5% combined incidence of complications we report, and even the 23.0% rate at high-volume centers, is higher than usually recorded in single-center series (19, 30, 31, 32). Many coded events we classified as complications were either DI or other fluid and electrolyte abnormalities, which likely resolved soon after surgery. Although permanent DI—1–2% of patients treated in specialized centers (19, 31)—is a more meaningful endpoint, both DI and fluid and electrolyte abnormalities coded before discharge were associated with higher rates of death and adverse discharge disposition (Table 3). Conversely, because iatrogenic hypopituitarism is assessed after discharge, our reported rate (0.8%) is probably an underestimate, and could reflect the results of prior surgery in some cases. We found no correlation between coded iatrogenic hypopituitarism and provider caseload, mortality rates, or discharge disposition.
Another difficulty that affects studies using administrative databases to measure postoperative outcomes is that some adverse events also represent surgical indications, and the NIS database does not distinguish between the two. We were unable to study postoperative increase in visual loss for this reason, although these patients should have been coded as neurological complications of surgery.
Demonstrations of better outcomes after complex surgical procedures performed at higher-volume centers often lead to a call for concentration of care in a limited number of hands, as by regionalizing care or by requiring special certification for surgeons (37, 38, 39). Our study’s limitations should be carefully weighed before our conclusions are accepted as supporting regionalization of transsphenoidal pituitary surgery.
Administrative databases are accurate sources for "hard" endpoints such as mortality rates and discharge disposition, but coding of comorbidities and complications is known to be incomplete (40, 41, 42, 43, 44, 45). The low rates we report for acromegaly and Cushing’s disease in comparison to other registry-based studies (46, 47, 48, 49, 50, 51) probably indicate undercoding in the NIS, reducing sensitivity to the effects such factors have on outcome (42). (Some subpopulations are probably also overrepresented at tertiary centers, typically the source of large case series.) Although incomplete coding of complications, overcoding of risk factors, or misclassification of other, low-risk operations as transsphenoidal pituitary resections at high-volume centers could account for our results, such large-scale biases appear unlikely. Less aggressive resections at high-volume centers could cause the effect we observed. Specialist surgeons, however, uniformly advocate the most aggressive adenoma resection consistent with patient safety, and the higher endocrine cure rates they achieve (23, 25, 27, 33, 35) indicate that this goal is typically realized.
Our study assessed only short-term endpoints because of the limitations of our data source. Because many patients discharged to short-term rehabilitation centers return home when recovery is complete, using short-term outcomes exaggerates the magnitude of the difference in long-term functional outcomes (and other potentially transient adverse outcomes such as DI) between high- and low-volume hospitals. Some important long-term outcomes, such as cure of Cushing’s disease or acromegaly and relief of preoperative visual loss, could not be assessed for the same reason. Our surgeon-caseload analyses do not account for the participation of two cosurgeons (a neurosurgeon and an otolaryngologist) in some transsphenoidal operations. The NIS identifies one primary surgeon per admission, whose specialty is not indicated. Whether the experience level of one or both surgeons, or of their combination as a team, affects outcome most has not yet been studied for any operation involving co-surgeons, including transsphenoidal surgery.
The most likely bias that could cause the volume-outcome effect we observed is the concentration of low-risk patients at high-volume centers. We found substantial evidence of this bias in our study: high-volume providers’ patients were younger, had admissions that were more often routine rather than urgent or emergent, and had less medical comorbidity, as well as higher socioeconomic status and better insurance—both likely to prompt more liberal use of screening studies for minor symptoms (and presentation with smaller tumors). Noniatrogenic panhypopituitarism and visual loss, both hallmarks of macroadenomas, were more common at low-volume centers. Cushing’s disease was the only unfavorable patient characteristic more common at high-volume centers.
Although we adjusted for these factors, adjustment for a risk index incorporating both clinical data (e.g. tumor size, history of prior transsphenoidal surgery, and preoperative endocrine and visual status) and the administrative data we used would likely reduce the magnitude of the volume-outcome effect we observed. A population-based study that incorporated clinical data would be valuable in confirming our findings.
Figures 1, 2, and 5 suggest that adverse outcomes become progressively less likely as provider caseload increases. Without clinical data on baseline risk or surgical efficacy, such as endocrine and visual cure, choosing a single, minimal acceptable caseload for providers (above which care is "optimal" and below which care is "unacceptable") might be misleading. Our results and studies linking more-experienced surgeons and higher endocrine cure rates (23, 25, 27, 33, 35) suggest that care will be optimized by choosing the most experienced surgeon and hospital available.
Ideally, a study on pituitary tumor surgery outcomes would describe surgical efficacy (probability and durability of endocrine and visual cure) as well as safety, but our data source lacked this information. Measures of surgical appropriateness at high- and low-volume centers would also be of interest. Sosa et al. (6) found that high-caseload surgeons had lower thresholds to perform parathyroidectomy than low-caseload surgeons; neither group followed consensus guidelines for the operation. A similar study on transsphenoidal pituitary surgery should be a priority for future research. A national or regional registry of pituitary tumor treatment outcomes, with input from both endocrinologists and surgeons, would facilitate such studies, as well as identification of factors leading to better outcomes at high-volume centers.
|
Conclusions |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionConclusionsReferences |
|---|
|
Footnotes |
|---|
Received March 17, 2003.
Accepted July 14, 2003.
|
References |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionConclusionsReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  O. M. Dekkers, N. R. Biermasz, A. M. Pereira, J. A. Romijn, and J. P. Vandenbroucke Mortality in Acromegaly: A Metaanalysis J. Clin. Endocrinol. Metab., January 1, 2008; 93(1): 61 - 67. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
O. M. Dekkers, N. R. Biermasz, A. M. Pereira, F. Roelfsema, M. O. van Aken, J. H. C. Voormolen, and J. A. Romijn Mortality in Patients Treated for Cushing's Disease Is Increased, Compared with Patients Treated for Nonfunctioning Pituitary Macroadenoma J. Clin. Endocrinol. Metab., March 1, 2007; 92(3): 976 - 981. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. W. Findling and H. Raff Cushing's Syndrome: Important Issues in Diagnosis and Management J. Clin. Endocrinol. Metab., October 1, 2006; 91(10): 3746 - 3753. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J M Hanson, H S Kooistra, J A Mol, E Teske, and B P Meij Plasma profiles of adrenocorticotropic hormone, cortisol, {alpha}-melanocyte-stimulating hormone, and growth hormone in dogs with pituitary-dependent hyperadrenocorticism before and after hypophysectomy. J. Endocrinol., September 1, 2006; 190(3): 601 - 609. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. P. Gillam, M. E. Molitch, G. Lombardi, and A. Colao Advances in the Treatment of Prolactinomas Endocr. Rev., August 1, 2006; 27(5): 485 - 534. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G. Minniti, V. Esposito, M. Piccirilli, A. Fratticci, A. Santoro, and M.-L. Jaffrain-Rea Diagnosis and management of pituitary tumours in the elderly: a review based on personal experience and evidence of literature Eur. J. Endocrinol., December 1, 2005; 153(6): 723 - 735. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
P. Caturegli, C. Newschaffer, A. Olivi, M. G. Pomper, P. C. Burger, and N. R. Rose Autoimmune Hypophysitis Endocr. Rev., August 1, 2005; 26(5): 599 - 614. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. Melmed, R. Sternberg, D. Cook, A. Klibanski, P. Chanson, V. Bonert, M. L. Vance, D. Rhew, D. Kleinberg, and A. Barkan A Critical Analysis of Pituitary Tumor Shrinkage during Primary Medical Therapy in Acromegaly J. Clin. Endocrinol. Metab., July 1, 2005; 90(7): 4405 - 4410. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 F. G. Barker II, W. T. Curry Jr., and B. S. Carter Surgery for primary supratentorial brain tumors in the United States, 1988 to 2000: The effect of provider caseload and centralization of care Neuro-oncol, January 1, 2005; 7(1): 49 - 63. [Abstract] [PDF]
|
|
| ||||||||||||||||||||
