Inpatient Complications After Transsphenoidal Surgery in Cushing s Versus Non-Cushing s Disease Patients

Transcription

1 595424AORXXX / Annals of Otology, Rhinology & LaryngologySvider et al research-article2015 Article Inpatient Complications After Transsphenoidal Surgery in Cushing s Versus Non-Cushing s Disease Patients Annals of Otology, Rhinology & Laryngology 2016, Vol. 125(1) 5 11 The Author(s) 2015 Reprints and permissions: sagepub.com/journalspermissions.nav DOI: / aor.sagepub.com Peter F. Svider, MD 1, Milap D. Raikundalia, BS 2, Morgan J. Pines, BS 2, Soly Baredes, MD, FACS 2,3, Adam J. Folbe, MD 1,4, James K. Liu, MD, FACS 2,3,5, and Jean Anderson Eloy, MD, FACS 2,3,5,6 Abstract Objective: Transsphenoidal surgery (TSS) harbors a potential for hypopituitarism, cerebrospinal fluid (CSF) leaks, and other complications. We utilized the Nationwide Inpatient Sample Database (NIS) to compare inpatient complication rates between Cushing s disease (CD) and non-cushing s disease (NCD) patients undergoing TSS. Methods: Inpatient hospitalization data for 960 CD and NCD patients who underwent TSS between 2002 and 2010 were accessed. Demographic information, outcomes, and complication rates were evaluated. Results: Patients with CD had a female predilection (81.7%) and were younger (40.5 ± 14.4 years) than NCD patients (47.8% female; 52.1 ± 16.3 years) (P <.001). Length of stay and total charges did not differ between groups. Patients with CD had significantly greater postoperative diabetes insipidus rates (14.0% vs 9.6%, P <.001) and urinary/renal complications (1.7% vs 0.9%, P =.027). After adjusting for possible confounders, the relationship between urinary/renal complications and CD status strengthened. There was no difference in rates of CSF leak and iatrogenic pituitary disorders overall. Conclusion: No differences were noted in the rate of early CSF leaks between postoperative TSS CD and NCD patients. Postoperative diabetes insipidus did not significantly differ between groups after adjusting for confounders. Only odds of urinary/renal complications in CD patients was significant after adjustment. Keywords Cushing s disease, Cushing s syndrome, transsphenoidal, benign pituitary neoplasm, Nationwide Inpatient Sample, cerebrospinal fluid leak, diabetes insipidus Introduction Significant advances in the management of Cushing s disease have ensued over the past century, including a resurrection of the transsphenoidal route for surgical intervention (TSS) initially developed (and subsequently abandoned) by Cushing in the early 1900s. 1 Surgery has largely become the standard of care in the initial treatment of Cushing s syndrome caused by the pituitary, 2-4 and further technological innovations have heralded a recent evolution from the microscopic technique to endoscopic approaches. 5 Despite a superior safety profile and decreased morbidity relative to transcranial approaches, TSS harbors risks, including a potential for hypopituitarism as well as cerebrospinal fluid (CSF) leak Some have suggested the latter risk is especially pronounced in Cushing s disease, characterized by poor wound healing due to a high cortisol environment. 12,13 Nonetheless, population-based studies definitively illustrating an increased CSF leak rate in Cushing s disease patients are lacking. There are no large cohort analyses comparing incidence of other complications between Cushing s patients and non-cushing s patients. Our primary objective was to examine the Nationwide Inpatient Sample Database 1 Department of Otolaryngology Head and Neck Surgery, Wayne State University School of Medicine, Detroit, Michigan, USA 2 Department of Otolaryngology Head and Neck Surgery, Rutgers New Jersey Medical School, Newark, New Jersey, USA 3 Center for Skull Base and Pituitary Surgery, Neurological Institute of New Jersey, Rutgers New Jersey Medical School, Newark, New Jersey, USA 4 Department of Neurosurgery, Wayne State University School of Medicine, Detroit, Michigan, USA 5 Department of Neurological Surgery, Rutgers New Jersey Medical School, Newark, New Jersey, USA 6 Department of Ophthalmology and Visual Science, Rutgers New Jersey Medical School, Newark, New Jersey, USA Corresponding Author: Jean Anderson Eloy, MD, FACS, Rutgers New Jersey Medical School, 90 Bergen Street, Suite 8100, Newark, NJ 07103, USA. jean.anderson.eloy@gmail.com

2 6 Annals of Otology, Rhinology & Laryngology 125(1) Table 1. ICD-9 Codes Utilized for Complications. Complication ICD-9 Code Postoperative neurological complications Pulmonary complications , Thromboembolic complications including deep venous , , , thrombosis and pulmonary embolism Cardiac complications ,997.1 Urinary and renal complications , Hemorrhage or hematoma complicating a procedure Diabetes insipidus Fluid and electrolyte abnormalities Cerebrospinal fluid rhinorrhea Iatrogenic pituitary disorders (NIS), a population-based resource offering short-term outcomes, length of stay, complications, and hospital charges, and evaluate whether patients undergoing TSS for Cushing s disease have a greater rate of CSF leak and other complications than patients with other pituitary tumors. Materials and Methods The authors evaluated the NIS database available from the Agency for Healthcare Research and Quality for information regarding inpatient hospitalizations of patients who had undergone TSS from 2002 to The NIS represents a 20% sample of all discharges from non-federal hospitals, with participating hospitals located in all 50 states. Rehabilitation and long-term acute care hospitals are not included in this database. Data from approximately 7 million hospital stays per year are included in this resource, representing greater than 36 million hospitalizations when weighted. 14 This nationally representative sample and its large sample size have proven invaluable in numerous prior analyses examining a wide range of topics. 11,15-24 Patients with a primary procedure of transsphenoidal surgery were included in the present analysis, and the Clinical Classification Software codes (CCS) and International Classification of Disease 9th revision (ICD-9) codes reported in Table 1 were used as patient inclusion criteria for examination of complications. These data and demographic information were organized and compared between patients who had undergone TSS for Cushing s disease versus non- Cushing benign pituitary lesions. Specifically, the primary diagnosis code (benign pituitary neoplasm) was used in conjunction with primary procedure codes (biopsy of pituitary gland, transsphenoidal approach), (partial excision of pituitary gland, transsphenoidal approach), and (total excision of pituitary gland, transsphenoidal approach). The comorbidity code (Cushing s syndrome) was utilized. We classified patients with the diagnosis of Cushing s syndrome undergoing TSS as Cushing s disease for the purpose of this analysis as the NIS does not have a specific code for Cushing s disease. Data collection was completed in October Statistical Analysis Continuous variables were compared using an independent t test, while categorical variables were compared using Pearson s chi-square analysis. Bivariate and multivariate regression was performed to assess for confounders. Threshold for significance was set at P <.05. SPSS version 22 (IBM, Armonk, New York, USA) was used for statistical calculation. Results Using the search criteria outlined, there were 13, 070 patients included in the NIS who underwent transsphenoidal resections from 2002 to 2010, including 12, 110 patients without Cushing s disease and 960 patients with Cushing s disease undergoing TSS. There was a temporal increase in the amount of TSS performed in both cohorts over this time period (Figure 1). While non-cushing s patients had a relatively even gender distribution, there was a considerable female predilection among the Cushing s cohort, as only 17.1% of Cushing s disease patients were men (Table 2). Cushing s disease patients were statistically younger, while no differences were observed upon comparison of length of stay and total charges (Table 2). Upon further analysis by age ranges, there was a statistically greater proportion of patients younger than 40 among the Cushing s cohort than the non-cushing s cohort (51.5% vs 24.3%, P <.001) (Table 2). There were significantly greater proportions of non-cushing s patients within all remaining older age groups (Table 2) (P values <.05). Significant differences in racial distribution were appreciated between the Cushing s and non-cushing s cohorts. There was a greater proportion of white patients (57.0% vs 46.7%, P <.001)

3 Svider et al 7 Figure 1. Temporal trends in transsphenoidal surgery among the cohorts analyzed. X-axis depicts years. and lesser proportion of black patients (6.5% vs 12.7%, P <.001) in the Cushing s cohort (Table 2). There was a significantly greater proportion of obesity within the Cushing s cohort versus the non-cushing s cohort (24.7% vs 8.7%, P <.001) (Table 2). Upon comparison of complication rates between these groups, statistical differences were noted with regards to urinary and renal complications, hemorrhage/hematoma, and diabetes insipidus (Table 3). Specifically, there was a higher rate of urinary and renal complications as well as diabetes insipidus in patients with Cushing s disease, while there was a lower rate of perioperative hemorrhage/hematoma in patients with Cushing s disease. There were no statistically significant differences in mortality, the overall presence of iatrogenic pituitary disorders, and rates of CSF rhinorrhea upon comparison between these groups (P values >.05) (Table 3). Rates for other complications are detailed in Table 3. Because of the significantly higher number of obese patients in the Cushing s disease group, further analysis was performed to determine whether this obesity had an effect on CSF rhinorrhea; obesity did not appear to have a significant impact on CSF rhinorrhea when examining the non-cushing and Cushing cohorts individually (Table 4). Logistic regression analysis was performed for each complication that differed between the 2 cohorts, namely, urinary and renal, hemorrhage or hematoma, and diabetes insipidus. The model was adjusted based on the significant characteristic differences observed between the Cushing s and non-cushing s cohorts in Table 2. Unadjusted odds reflected similar relationships seen in the chi-square analysis; Cushing s cohort experienced greater urinary and renal complications and diabetes insipidus but fewer hemorrhage or hematoma complications, all of which reached statistical significance (Table 5). After adjusting for age, sex, race, and obesity status, only urinary and renal complications retained significance. Furthermore, the unadjusted odds (1.80; P =.029) were substantially less than the adjusted odds (4.68; P <.001) (Table 5). The accuracy of urinary and renal complications regression models were evaluated using Hosmer and Lemeshow test, resulting in P values of nd.496 for unadjusted and adjusted, respectively, suggesting the latter to be the better fitting model. Nagelkerke R 2 were and for unadjusted and adjusted, respectively. Discussion A prior study examining inpatient hospitalizations of postoperative TSS Cushing s disease patients noted that length of stay, hospital and surgeon volume, and postoperative complications had predictive value with regards to

4 8 Annals of Otology, Rhinology & Laryngology 125(1) Table 2. Characteristics of Cushing s Versus Non-Cushing s Patient Cohort. Characteristics Cushing s (N=960), (%) Non-Cushing s (N=12,110), (%) P Value Mean age, y 40.5 ± ± 16.3 <.001 Age cohorts, % < <.001 > <.001 Unknown Sex, % Male <.001 Female <.001 Unknown Race, % White <.001 Black <.001 Hispanic Asian/Pacific Islander Other Unknown Obesity status, % Obese <.001 Non-obese <.001 Unknown Mean length of stay, d Mean hospital charge, dollars 46, , a Analysis between groups of a given variable (age, sex, race, obesity status). Table 3. Postoperative Complications in Cushing s Versus non-cushing s Patients. Complications Cushing s N = 960, (%) Non-Cushing s N = 12, 110, (%) P Value Neurological Pulmonary Cardiac Thromboembolic Urinary and renal Hemorrhage or hematoma Diabetes insipidus <.001 Fluid and electrolyte abnormalities Cerebrospinal fluid rhinorrhea Iatrogenic pituitary disorders Mortality resource utilization. 25 Patil and colleagues 11 also examined the NIS database for characteristics of patients who had undergone TSS for Cushing s disease from 1993 to 2002, noting a 1.4% rate of CSF leak. Since there have been technological advances and other evolving trends in TSS since that time period, 5 the present analysis serves as a valuable update. Additionally, we evaluated a control non-cushing s cohort for comparison in our analysis, allowing us to evaluate whether there are higher rates of CSF leak, endocrinopathies, and other complications in Cushing s disease patients. Nemergut et al 26 retrospectively examined perioperative records of patients who had undergone TSS between 1995 and 2001 at their institution for factors associated with diabetes insipidus. Among this cohort, 183 patients underwent surgical intervention for Cushing s disease. The authors noted an increased diabetes insipidus risk relative to those with other pituitary lesions within this cohort as well as

5 Svider et al 9 Table 4. Impact of Obesity on Cerebrospinal Fluid Rhinorrhea. Cohort Obese Non-Obese P Value Odds Ratio 95% Confidence Interval Cushing s, % Non-Cushing s, % Table 5. Logistic Regression Analysis of Associations of Post-Transsphenoidal Surgery Complications to Cushing s Status. a Complications Unadjusted Odds P Value Adjusted Odds P Value Urinary and renal <.001 Hemorrhage or hematoma Diabetes insipidus 1.53 < a Odds ratios reflect Cushing s patients relative to non-cushing s patients. Adjusted odds ratios accounted for age, sex, race, and obesity. among patients who underwent TSS for craniopharyngiomas and Rathke s cleft cysts. The present analysis does not support the findings of this prior single institution cohort. Our analysis encompasses a nationally representative sample of 13, 070 TSS patients, including 960 Cushing s disease patients. Initially, our Pearson s chi-square and crude regression analysis revealed a possible increased proportion of diabetes insipidus events in the Cushing s disease cohort. However, unlike the aforementioned study, a logistic regression analysis was performed to adjust for age, sex, race, and obesity status. The adjusted model revealed no substantial differences in the proportion of diabetes insipidus events between Cushing s and non-cushing s disease. However, the overall proportion of patients experiencing diabetes insipidus in both groups necessitates the inclusion of this potential complication in a comprehensive preoperative informed consent process. These findings reinforce the importance of operating surgeons being mindful of diabetes insipidus postoperatively for TSS and having an equal threshold for aggressively managing this condition regardless of Cushing s disease status. Cerebrospinal fluid leak rates in TSS patients have been previously reported in the literature, with estimates varying depending on the source. A meta-analysis encompassing publications from 1998 through 2010 noted leak rates of 5% to 7% depending on whether microscopic or endoscopic approaches were employed. 27 Utilizing cohorts in California and Florida all-payer databases from more recent patient data ( ), Krings and colleagues 28 noted an even lower rate of skull base complications at 3.6%, with this figure covering both CSF leaks and postoperative bacterial meningitis. Our figures for both Cushing s and non-cushing s disease patients were slightly lower, although they were consistent with other analyses utilizing the NIS. 29 A potential reason for these lower figures, and thus a limitation of our analysis, could have been that this database covers only the actual hospitalization and may not account for delayed presentations of CSF leaks. We do not believe this would have significantly affected results as most CSF leaks present in the first few days following surgery. Nonetheless, this is the first population-based analysis demonstrating there is not an increased CSF leak rate in the Cushing s disease population in the immediate postoperative period when compared with non-cushing s disease TSS patients. As CSF leaks are a relatively uncommon but potentially serious issue, obtaining large sample sizes from clinical cohorts would be very difficult, necessitating the use of population-based databases such as the NIS for such an analysis. Furthermore, as the NIS constitutes approximately one-fifth of all inpatient admissions to US hospitals, this is likely an accurate representation of CSF leak rates in TSS patients. Elevated body mass index (BMI) has been previously associated with a greater risk of postoperative CSF rhinorrhea following TSS. 7 Dlouhy and colleagues 7 performed a retrospective review of 121 patients who underwent TSS, noting that the average BMI of patients with CSF rhinorrhea was greater than those with no postoperative leak. In our analysis, we noted a significantly greater proportion of obesity, as defined by BMI, among Cushing s disease patients than in the non-cushing s disease cohort. Nonetheless, we found no difference in CSF rhinorrhea between the Cushing s disease and non-cushing s disease cohorts and further noted no effect of obesity on our CSF leak results (Table 4), indicating this was not a confounding factor impacting our outcomes. Although we could postulate that the minimally invasive nature of TSS may preclude the deleterious effects that obesity may confer to CSF rhinorrhea or that surgeons performing these procedures in obese patients or patient with Cushing s disease are being more careful due to the historical risk of postoperative CSF leakage associated with these conditions (and therefore decreasing the occurrence of this complication in these patients), these notions would be purely speculative. Hence,

6 10 Annals of Otology, Rhinology & Laryngology 125(1) no definite explanation could be given for these findings based on this analysis. No significant differences were noted in the occurrence of iatrogenic pituitary disorders postoperatively upon comparison of the Cushing s and non-cushing s disease cohorts. Although hemorrhage and hematoma complications appeared to afflict a greater proportion of non-cushing s disease patients, this relationship did not hold true after adjustment for other covariates. A significantly higher percentage of urinary and renal complications was noted postoperatively however (1.7% vs 0.9%; P =.027). Based on the regression analysis, the odds ratio not only more than doubled after adjustment but also increased in significance (1.80, P =.029 vs 4.68, P <.001). In addition, the adjusted model met the criteria for more accuracy compared to the unadjusted. We believe this provides strong evidence of negative confounding relationship within 1 of the covariates within the model. It is not clear which covariates in this model are confounders. Overall, this finding reinforces the strength of the association between urinary and renal complications in Cushing s disease patients. It must be emphasized that no causal relationship can be inferred from this analysis. In addition, the relatively small Nagelkerke value for the adjusted model suggests that although Cushing s disease has a noteworthy relationship with urinary and renal events compared to the other independent variables in this model, there may be other factors that make a large contribution to this complication that are outside the scope of this study. The specific reason for greater renal/urinary dysfunction is unclear and represents a further area for study. Nonetheless, this finding reinforces the importance of carefully evaluating renal function in the immediate postoperative period. Although this analysis represents a comprehensive and nationally representative comparison of the inpatient course of TSS patients with Cushing s disease versus other pituitary tumors, there are several limitations inherent to our study design. While utilizing the NIS has the advantages a multi-institutional resource offers over single-institution cohorts, including greater external validity, several specific clinical details that would have enhanced this analysis were not available. For example, the NIS database does not contain information on tumor size, classification of functional adenoma, extent of local invasion, or cortisol levels, all of which may influence postoperative complications in patients undergoing TSS for pituitary tumors. In addition, perioperative medical intervention and complication chronology cannot be extrapolated. There was also a large disparity in the number of cases of Cushing s disease compared to the non-cushing s disease group, partially due to our inclusion requirement for diagnosis of both benign pituitary lesion and Cushing s disease. This study is also subject to all the limitations inherent to all large population-based databases. Nonetheless, the large sample size and nationally representative nature of this resource permits us to examine and compare complication rates with appropriate external validity to allow our conclusions to be generalizable. Conclusions Transsphenoidal approaches result in a low rate of postoperative complications. Despite the theoretical concern that Cushing s disease patients have poorer wound healing due to higher circulating cortisol levels, no differences were noted in the rate of early CSF leaks between these patients and those undergoing TSS for other benign pituitary neoplasms. Overall, fewer than 2% of patients had CSF rhinorrhea during their hospitalization. While both non-cushing s patients and Cushing s disease patients had a sizeable rate of postoperative diabetes insipidus, this complication did not predominate in either group. This still reinforces the importance of practitioners being mindful of this potential issue in postoperative transsphenoidal patients. Urinary and renal complications were strongly associated with Cushing s disease patients, more so after adjustment for confounders. This necessitates greater care of Cushing s disease patients following TSS for these complications but also warrants further studies to elucidate the pathophysiological mechanism. Utilizing a nationally representative sample, findings from this analysis are generalizable, although unique institution-specific practices may have an impact on complications and outcomes. Declaration of Conflicting Interests The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article. Funding The author(s) received no financial support for the research, authorship, and/or publication of this article. References 1. Patel SK, Husain Q, Eloy JA, Couldwell WT, Liu JK. Norman Dott, Gerard Guiot, and Jules Hardy: key players in the resurrection and preservation of transsphenoidal surgery. Neurosurg Focus. 2012;33:E6. 2. Atkinson AB, Kennedy A, Wiggam MI, McCance DR, Sheridan B. Long-term remission rates after pituitary surgery for Cushing s disease: the need for long-term surveillance. Clin Endocrinol (Oxf). 2005;63: Chee GH, Mathias DB, James RA, Kendall-Taylor P. Transsphenoidal pituitary surgery in Cushing s disease: can we predict outcome? Clin Endocrinol (Oxf). 2001;54: Choudhry OJ, Choudhry AJ, Nunez EA, et al. Pituitary tumor apoplexy in patients with Cushing s disease: endocrinologic

7 Svider et al 11 and visual outcomes after transsphenoidal surgery. Pituitary. 2012;15: Svider PF, Keeley BR, Husain Q, et al. Regional disparities and practice patterns in surgical approaches to pituitary tumors in the United States. Int Forum Allergy Rhinol. 2013;3: Berker M, Hazer DB, Yucel T, et al. Complications of endoscopic surgery of the pituitary adenomas: analysis of 570 patients and review of the literature. Pituitary. 2012;15: Dlouhy BJ, Madhavan K, Clinger JD, et al. Elevated body mass index and risk of postoperative CSF leak following transsphenoidal surgery. J Neurosurg. 2012;116: Hofstetter CP, Shin BJ, Mubita L, et al. Endoscopic endonasal transsphenoidal surgery for functional pituitary adenomas. Neurosurg Focus. 2011;30:E Senior BA, Ebert CS, Bednarski KK, et al. Minimally invasive pituitary surgery. Laryngoscope. 2008;118: Wang F, Zhou T, Wei S, et al. Endoscopic endonasal transsphenoidal surgery of 1,166 pituitary adenomas. Surg Endosc. 2015;29(6): Patil CG, Lad SP, Harsh GR, Laws ER Jr, Boakye M. National trends, complications, and outcomes following transsphenoidal surgery for Cushing s disease from 1993 to Neurosurg Focus. 2007;23:E Zanation AM, Thorp BD, Parmar P, Harvey RJ. Reconstructive options for endoscopic skull base surgery. Otolaryngol Clin North Am. 2011;44: Zanation AM, Carrau RL, Snyderman CH, et al. Nasoseptal flap reconstruction of high flow intraoperative cerebral spinal fluid leaks during endoscopic skull base surgery. Am J Rhinol Allergy. 2009;23: Overview of the National (Nationwide) Inpatient Sample (NIS). Accessed October 5, Saeed F, Adil MM, Piracha BH, Qureshi AI. Outcomes of endovascular versus intravenous thrombolytic treatment for acute ischemic stroke in dialysis patients. Int J Artif Organs. 2014;37(10): Schoenfeld AJ, Wahlquist TC. Mortality, complication-risk and total charges following the treatment of epidural abscess. Spine J. 2015;15(2): Smith ER, Butler WE, Barker FG II. Is there a July phenomenon in pediatric neurosurgery at teaching hospitals? J Neurosurg. 2006;105: Terry AR, Barker FG II, Leffert L, Bateman BT, Souter I, Plotkin SR. Neurofibromatosis type 1 and pregnancy complications: a population-based study. Am J Obstet Gynecol. 2013;209:e41-e Wang HH, Wiener JS, Ferrandino MN, Lipkin ME, Routh JC. Complications in surgical management of upper tract calculi in spina bifida patients: analysis of nationwide data. J Urol. 2015;193(4): Chung TK, Rosenthal EL, Magnuson JS, Carroll WR. Transoral robotic surgery for oropharyngeal and tongue cancer in the United States. Laryngoscope. 2015;125(1): Ishman SL, Ishii LE, Gourin CG. Temporal trends in sleep apnea surgery: Laryngoscope. 2014;124: Ouyang D, El-Sayed IH, Yom SS. National trends in surgery for sinonasal malignancy and the effect of hospital volume on short-term outcomes. Laryngoscope. 2014;124: Richmon JD, Quon H, Gourin CG. The effect of transoral robotic surgery on short-term outcomes and cost of care after oropharyngeal cancer surgery. Laryngoscope. 2014;124: Villwock JA, Jones K. Outcomes of early versus late tracheostomy: Laryngoscope. 2014;124: Little AS, Chapple K. Predictors of resource utilization in transsphenoidal surgery for Cushing disease. J Neurosurg. 2013;119: Nemergut EC, Zuo Z, Jane JA Jr, Laws ER Jr. Predictors of diabetes insipidus after transsphenoidal surgery: a review of 881 patients. J Neurosurg. 2005;103: DeKlotz TR, Chia SH, Lu W, Makambi KH, Aulisi E, Deeb Z. Meta-analysis of endoscopic versus sublabial pituitary surgery. Laryngoscope. 2012;122: Krings JG, Kallogjeri D, Wineland A, Nepple KG, Piccirillo JF, Getz AE. Complications following primary and revision transsphenoidal surgeries for pituitary tumors. Laryngoscope. 2015;125(2): Barker FG II, Klibanski A, Swearingen B. Transsphenoidal surgery for pituitary tumors in the United States, : mortality, morbidity, and the effects of hospital and surgeon volume. J Clin Endocrinol Metab. 2003;88: