Predictors of diabetes insipidus after transsphenoidal surgery: a review of 881 patients

Transcription

1 J Neurosurg 103: , 2005 Predictors of diabetes insipidus after transsphenoidal surgery: a review of 881 patients EDWARD C. NEMERGUT, M.D., ZHIYI ZUO, M.D., PH.D., JOHN A. JANE JR., M.D., AND EDWARD R. LAWS JR., M.D. Departments of Anesthesiology, Medicine, Neurosurgery, and Pediatrics, University of Virginia Health System, Charlottesville, Virginia Object. Diabetes insipidus (DI) is a common complication of transsphenoidal surgery. The purpose of this study was to elucidate patient- and surgery-specific risk factors for DI. Methods. The perioperative records of 881 patients who had undergone transsphenoidal microsurgery at the authors institution between January 1995 and June 2001 were reviewed. Among 857 patients without preoperative DI, the overall incidence of immediate postoperative DI was 18.3%, with 12.4% of patients requiring treatment with desmopressin at some point during their hospitalization. Persistent DI requiring long-term treatment with desmopressin was noted in 2% of all patients. An observable intraoperative cerebrospinal fluid (CSF) leak was strongly associated with an increased incidence of both transient (33.3%) and persistent (4.4%) DI. Craniopharyngioma and Rathke cleft cyst (RCC) were also associated with an increased incidence of transient and persistent DI, whereas repeated operation was not. Among patients with pituitary adenomas, those with Cushing s disease had an increased risk of transient (22.2%), but not persistent, DI. Patients with a microadenoma were more likely to suffer transient DI than those harboring a macroadenoma (21.6 compared with 14.3%) but were not more likely to experience persistent DI. Conclusions. Diabetes insipidus remains a common complication of transsphenoidal surgery; however, it is most frequently transient in nature. Patients with an intraoperative CSF leak, a microadenoma, a craniopharyngioma, or an RCC appear to have an increased risk of transient DI. Risk factors for persistent DI include an intraoperative CSF leak, a craniopharyngioma, or an RCC. KEY WORDS diabetes insipidus pituitary adenoma craniopharyngioma Rathke cleft cyst endocrinopathy pituitary surgery D Abbreviations used in this paper: ADH = antidiuretic hormone; CD = Cushing s disease; CI = confidence interval; CSF = cerebrospinal fluid; DI = diabetes insipidus; MR = magnetic resonance; OR = odds ratio; RCC = Rathke cleft cyst; UVA = University of Virginia. 448 ISORDERS of water balance caused by disturbances in ADH secretion and posterior pituitary function remain a common cause of morbidity among patients undergoing transsphenoidal surgery. Indeed, abnormalities of ADH secretion resulting in postoperative central DI and the syndrome of inappropriate secretion of ADH are the most common early postoperative endocrine complications. 30 Diabetes insipidus may complicate the postoperative course in as many as 30% of patients. 4,14,30 Although the disease is transient and benign in the overwhelming majority of cases, prolonged or permanent DI, reportedly resulting from more proximal damage to the pituitary stalk and cell bodies in the hypothalamic nuclei, may also occur. 2,27,28 Some patients with DI will present with a triphasic pattern of water imbalance as described in the classic work of Fisher and Ingram. 10 Despite the ubiquitous nature of DI following transsphenoidal surgery, minimal detailed information regarding patient- and surgery-specific risk factors has been elucidated. Thus, we chose to review this specific population to better characterize patient and surgical risk. This retrospective review encompasses data from 881 patients who underwent transsphenoidal microsurgery performed by a single surgeon. Clinical Material and Methods After project approval from the UVA human investigations committee, we established a patient database compliant with the Health Insurance Portability and Accountability Act of We reviewed the perioperative records of 881 patients who had undergone transsphenoidal microsurgery at the UVA Medical Center between January 1995 and June A single surgeon (E.R.L.) performed each procedure, normally using an endonasal transsphenoidal procedure. 18 A transseptal sublabial approach was seldom performed as it is generally reserved for children or patients with very large tumors. 17,18 Endonasal endoscopy was not used for tumor resection in any patient in this series. Intraoperative fluoroscopy or computer-guided frameless stereotaxy 8,18 were used to guide surgical removal. After resection of the adenoma, the sellar floor was reconstructed using septal bone as previously described. 18 If a readily observable CSF leak was noted, a fat graft was obtained from the patient s abdomen and placed within the sella prior to reconstruction of the sellar floor. During surgery and the immediate postoperative period,

2 Diabetes insipidus and transsphenoidal surgery TABLE 1 Clinical disease and pathological diagnosis in 881 patients* TABLE 2 Adenoma size and the diagnosis and treament of DI Clinical Disease No. of Patients Adenoma Size nonfunctioning adenoma 348 CD 183 acromegaly 121 prolactinoma 87 craniopharyngioma 39 RCC 33 arachnoid cyst 11 clival chordoma 9 meningioma 8 LH/FSH secreting lesion 8 CSF leak 5 metastatic carcinoma 4 colloid cyst 3 epidermoid cyst 3 lymphoma 3 thyrotropic (TSH-secreting) lesion 3 plasma cell tumor 2 other 11 total 881 * FSH = follicle-stimulating hormone; LH = luteinizing hormone; TSH = thyroid-stimulating hormone. all patients received unrestricted intravenous fluids. All patients also received intravenous and/or inhaled anesthetic agents as well as opioids and muscle relaxants; however, carbamazepine, mannitol, or other diuretics were not routinely administered. At the UVA Medical Center, a protocol for monitoring fluid balance and detecting DI during the postoperative period has been well established. Briefly, patients are weighed daily (at the same time, using the same scale) and the specific gravity of their urine is checked every 4 hours. Although Foley catheters are not routinely inserted, fluid intake and output are carefully recorded, and patients are questioned regarding thirst. Blood glucose, serum electrolytes (including sodium, potassium, and calcium), and serum osmolarities are determined daily. Serum sodium and osmolarity may be evaluated more frequently as indicated. Hyperglycemia is always excluded as a cause of polyuria and hyponatremia. For the purpose of this review, a patient was considered to have clinical evidence of DI if the diagnosis was noted on the discharge summary by the primary neuroendocrine team and if the patient had documented voluminous urine output ( 300 ml/hour) for more than 3 hours, with a specific gravity of less than A rising serum sodium level is considered to be evidence of DI but is not required to make the diagnosis or to initiate treatment. Given that the majority of patients in the postoperative period following transsphenoidal surgery are awake and alert and have intact thirst mechanisms as well as adequate access to fluids, desmopressin is not routinely administered. Nevertheless, desmopressin treatment in patients with the diagnosis of DI was recorded in the database. Follow-up records from both the neurosurgery and endocrinology clinics (if available) were reviewed to determine whether any patient had persistent DI and the need for long-term treatment with desmopressin. Finally, any patient with preexisting DI was excluded from the review. Statistical analysis was performed using Parameter Macroadenoma Microadenoma Total clinical diagnosis of DI* w/ DI w/o DI total rate of DI (%) in-hospital treatment w/ desmopressin treated w/ desmopressin no desmopressin total rate of desmopressin treatment (%) persistent DI requiring desmopressin w/ DI w/o DI total rate of DI (%) * Comparing patients with macroadenoma and those with microadenoma, p = for the clinical diagnosis of DI. Comparing patients with macroadenoma and those with microadenoma, p = for in-hospital desmopressin treatment of DI. Excludes three patients with incomplete follow up who had received inhospital desmopressin treatment of DI (two with macroadenomas and one with a microadenoma). Comparing adenoma size, p for persistent DI. the z-test. A probability level less than 0.05 was considered significant. Results The clinical disease and pathological diagnosis for each patient in this review is presented in Table 1. Of the 881 patients reviewed, 24 had preexisting DI and were excluded from the study. Of the remaining 857 patients, 157 (18.3%) had clinical evidence of DI during the immediate postoperative period. Diabetes insipidus was treated with desmopressin in 106 patients (12.4%) at some point during their hospitalization. Persistent DI requiring long-term treatment with desmopressin was noted in 17 patients (2%). Among the 743 patients who had a pituitary adenoma and no preoperative DI, 516 (69.4%) had macroadenomas. Microadenomas were found in 227 patients (30.6%). Diabetes insipidus was more likely to be diagnosed in patients with microadenomas than in those with macroadenomas (21.6 compared with 14.3%, respectively; p = 0.018) and more likely to be treated with desmopressin (14.5 compared with 9.3%, respectively; p = 0.049). Adenoma size had no impact on the development of persistent DI (Table 2). Of the 857 patients without preoperative DI, 186 presented for repeated transsphenoidal procedures. A repeated operation did not increase the incidence of either postoperative DI (21% in patients who underwent repeated surgery compared with 17.6% in patients who did not, p = 0.34) or disease necessitating treatment with desmopressin (12.9% in patients with repeated surgery compared with 12.2% in 449

3 E. C. Nemergut, et al. TABLE 3 Repeated operations and the clinical diagnosis and treatment of DI Repeated Op Parameter Yes No Total clinical diagnosis of DI w/ DI w/o DI total rate of DI (%) in-hospital desmopressin treatment treated w/ desmopressin no desmopressin total rate of desmopressin treatment (%) persistent DI requiring desmopressin treatment w/ DI w/o DI total rate of DI (%) * There was no significant difference in any of the comparisons. Excludes three patients who underwent repeated surgery and had an incomplete follow up (all had in-hospital desmopressin treatment). those without, p = 0.896). Furthermore, a repeated operation did not increase the incidence of persistent DI requiring long-term treatment with desmopressin (Table 3). An intraoperative CSF leak was noted in 273 patients (31.8%) and was associated with an increased incidence of DI (33.3% in patients with a leak compared with 11.3% in those without, p 0.001) as well as an increased need for treatment with desmopressin (22.7% in patients with a CSF leak compared with 7.5% in those without, p 0.001). Patients with a CSF leak were also noted to have an increased incidence of persistent DI (4.4 compared with 0.9% in patients without, p 0.002) requiring long-term desmopressin treatment. These data are summarized in Table 4. Among the subgroup of patients with a pituitary adenoma (743 patients), a CSF leak was also associated with an increased incidence of DI (31.7% in patients with a leak compared with 10% in those without, p 0.001) and an increased need for desmopressin treatment (20 compared with 6.9%, respectively; p 0.001). Only four patients with a pituitary adenoma suffered persistent DI requiring long-term treatment with desmopressin. Three of these four patients had an intraoperative CSF leak, but this difference was not statistically significant. Considering clinical disease and pathological diagnosis, patients presenting with a craniopharyngioma were more likely to have DI (62.1% in patients with craniopharyngioma compared with 18.3% in the entire study cohort, p 0.001) and to be treated with desmopressin (58.6 compared with 12.4%, respectively; p 0.001). Patients with a craniopharyngioma were also more likely to have persistent DI (31 compared with 2%, respectively; p 0.001) requiring TABLE 4 Intraoperative CSF leak and the diagnosis and treament of DI long-term desmopressin treatment. Although patients presenting with an RCC were more likely to have DI (38.7 compared with 18.3% in the entire study cohort, p 0.009), in-hospital treatment with desmopressin failed to reach statistical significance (19.4 compared with 12.4% in the entire study cohort, p 0.38). Nevertheless, patients with an RCC were more likely to have persistent DI (9.7 compared with 2% in the entire study cohort, p 0.027) requiring long-term treatment with desmopressin. In a comparison of the entire cohort of 857 patients, we found that the specific type of pituitary adenoma does not appear to impact the incidence of DI. These data are summarized in Table 5. The probability values, ORs, and 95% CIs were calculated based on a comparison between each disease group and the entire study cohort. Among patients with a pituitary adenoma (743 patients), the type of pituitary tumor only affected the incidence of DI if the patient had CD; that is, patients with CD (180 patients) were more likely to have DI than other patients with pituitary adenomas (22.2 compared with 14.7%, respectively; p = 0.026) and more likely to be treated with desmopressin (16.1 compared with 9.2%, respectively; p = 0.013). This finding corresponds to an OR of 1.65 (95% CI ) for the diagnosis of DI and an OR of 1.89 (95% CI ) for desmopressin treatment. No effect on the incidence of long-term DI requiring persistent desmopressin treatment was seen in patients with CD. Discussion Fat Graft Parameter Yes No Total clinical diagnosis of DI w/ DI w/o DI total rate of DI (%) in-hospital desmopressin treatment treated w/ desmopressin no desmopressin total rate of desmopressin treatment (%) persistent DI requiring desmopressin treatment w/ DI w/o DI total rate of DI (%) * Comparison between patients with and without intraoperative CSF leak. Excludes three patients with incomplete follow up (all had fat graft and in-hospital DI). Antidiuretic hormone is a nonapeptide that is synthesized in the hypothalamic paraventricular and supraoptic nuclei. 450

4 Diabetes insipidus and transsphenoidal surgery TABLE 5 Clinical disease and the diagnosis and treatment of DI* Disease Parameter ACR CPG CD NFA PRLA RCC Total clinical diagnosis of DI total no. of cases cases w/ transient DI overall incidence (%) p value NA OR NA 95% CI NA in-hospital desmopressin treatment total no. of cases cases treated w/ desmopressin overall incidence (%) p value NA OR NA 95% CI NA persistent DI requiring desmopressin treatment total no. of cases cases of persistent DI overall incidence (%) p value NA OR NA 95% CI NA NA NA * ACR = acromegaly; CPG = craniopharyngioma; NA = not applicable; NFA = nonfunctioning adenoma; PRLA = prolactinoma. Excludes three patients lost to follow up (two cases of acromegaly and one case of NFA). Not applicable because there was no occurrence of persistent DI. After its initial synthesis, the precursor hormone is transferred down axons through the hypothalamoneurohypophysial tract to the posterior pituitary. In the posterior pituitary, ADH undergoes final maturation to active hormone and is stored for future release. 30 Secretion of ADH is regulated primarily by plasma osmolarity, but other factors such as left atrial distention, circulating blood volume, exercise, and certain emotional states can also alter ADH release. The relative or absolute deficiency in ADH results in DI. After transsphenoidal surgery, DI can result from an interruption in the transport of ADH from the hypothalamus, impairment of ADH release from the posterior pituitary, or retrograde damage to the cell bodies in the hypothalamic nuclei. Thus, damage anywhere along the hypothalamus pituitary axis can result in DI. Our data revealed the overall incidence of transient DI to be 18.3%. If we include only patients who underwent surgery for removal of a pituitary adenoma, the incidence decreases to 16.6%. Approximately 67.5% of patients with clinical evidence of DI were treated at least once with desmopressin; that is, 12.4% of all patients and 10.9% of patients with pituitary tumors were treated with desmopressin at some point during hospitalization. In a series of 1571 patients with pituitary adenomas, Hensen, et al., 14 reported an overall incidence of 31% for hypotonic polyuria, with approximately 80% of those patients receiving desmopressin. Postoperatively, 24% of all patients who had undergone transsphenoidal surgery were treated with desmopressin at least once. We are unable to explain completely the approximately twofold decrease in the incidence of DI and treatment with desmopressin observed in our series. It has been previously proposed that the incidence of DI postsurgery is inversely related to the experience of the surgeon; 4 however, the procedures in both studies were clearly performed by very experienced surgeons. Given that Hensen and colleagues reported data from December 1982 to December 1995 and that we reported those from 1995 to 2001, it is possible that the apparent decrease in the incidence of DI results from improvements in surgical technique from 1982 to 2001, including improvements in radiographic imaging, such as MR imaging. Nevertheless, there is considerable variation regarding the reported incidence of transient DI. 4,14,19,23,25,26,29,30,32 The retrospective nature of most studies, including the present series, can confound results. Indeed, it is more likely that the apparent difference between our study and the one conducted by Hensen and associates 14 relates to differences in the definition of DI and the willingness to treat the disease with desmopressin. Perhaps the philosophy for initiating medical therapy is simply more strict at UVA. Given that the majority of postoperative patients are awake and alert and have intact thirst mechanisms as well as adequate access to fluids, the occurrence of significant volume depletion and hyperosmolarity is relatively uncommon. Consequently, desmopressin is administered only if there is a significant discrepancy in fluid intake and output, a rising serum sodium level ( 145 meq/l), and an excessive urine output that significantly interferes with sleep. Indeed, our reluctance to treat early DI is based on our recognition that postoperative DI is most often transient and self-limited. Nevertheless, different centers may have widely varying criteria for the treatment of DI with desmopressin and thus comparisons may be difficult. In contrast to the tremendous variation in the reported incidence of transient DI, most authors note fairly consistent results regarding the incidence of persistent DI, usually be- 451

5 E. C. Nemergut, et al. tween 0.5 and 1.5%. 2,4,14,29,30,32 In the present series, the overall incidence of persistent DI was 2% among all patients. When considering only those patients who had undergone surgery for removal of a pituitary adenoma, the incidence decreased to 0.68%. Given the extremely low incidence of persistent DI, our data underscore the fact that most DI is transient and self-limited. Practitioners should prevent the overtreatment of polyuria with desmopressin, especially in patients who can keep up with urinary losses through oral fluid intake. Excessive desmopressin treatment can result in hyponatremia and significant morbidity. Interestingly, transient DI and consequent desmopressin treatment are significantly more common after transsphenoidal surgery in patients harboring microadenomas (Table 2). Some authors have shown that DI is more common after the resection of macroadenomas. 2,14,23 Olson and colleagues 23 have asserted that more extensive exploration of the gland and stalk may be required for successful resection of microadenomas, especially when a tumor is very small or does not appear on preoperative imaging. Increased manipulation of the stalk and gland may result in more damage, however, and thus an increased risk of DI. Ciric and associates 4 have also posited that DI following the resection of a microadenoma is commonly the result of stalk manipulations. In patients with a macroadenoma, the tumor effectively displaces the posterior pituitary and the stalk upward, often leaving it covered with a fibrous layer. 4,23 This location and protective layer may protect the posterior pituitary and may further explain the increased incidence of DI after resection of a microadenoma. It has also been posited that new routes and/or mechanisms of ADH delivery may have developed in patients with macroadenoma. 19 The loss of normal posterior pituitary tissue by an expanding intrasellar mass is well recognized by neurosurgeons. Indeed, the loss of lipid-laden pituicytes can appear as a loss of normal high signal intensity on T 1 -weighted MR images. 6 This is particularly apparent in patients with macroadenomas. Such patients may be less susceptible to stalk manipulations and aberrations in ADH secretion. In fact, surgery may actually relieve the chronic distortion of the neurohypophysial stalk, leading to a surge in ADH release. 19 A repeated operation did not result in the increased incidence of transient or persistent DI. We had originally hypothesized that multiple operations involving more manipulations of the stalk and posterior pituitary might cause more damage and an increased incidence of DI; however, this obviously did not occur. A carefully conducted repeated operation is no more likely to precipitate DI than a carefully conducted resection of an untreated tumor. Furthermore, it is important to remember that total hypophysectomy does not necessarily lead to permanent DI. 3 Thus, even when there is only a small amount of residual gland present after several resections, DI might not occur. Another interesting finding is the high incidence of transient and persistent DI in patients who had an observable intraoperative CSF leak. Indeed, aside from having a craniopharyngioma, there is no other perioperative factor associated with a higher risk of DI. Although it is most likely that an intraoperative CSF leak may simply be a marker for a more aggressive resection involving more stalk manipulation, there may be other explanations. All patients in this series with an observable intraoperative CSF leak underwent placement of a fat graft. After fat is harvested from the patient s abdomen, it is bathed in chloramphenicol and swiped across cotton such that very small, yet visible, wisps of cotton cling to the fat. 18 The cotton is thought to engender a mild foreign body reaction that helps seal the leak. 18 In addition to helping seal the leak, this inflammatory process may also promote an increased risk of transient DI. Considering that other inflammatory processes such as meningitis 11 (including neonatal meningitis 5 ), abscess, 12 and tuberculosis 16 can all present with DI, the inflammatory process initiated by cotton and fat may directly result in an increased incidence of, at least, transient DI. In addition, perhaps chloramphenicol leaches from the fat graft and induces transient DI. Note, however, that there are absolutely no data to support a relationship between intrathecal chloramphenicol and DI. Regardless, a prospective study focused on the direct contribution of a fat graft to the development of DI would no doubt be interesting. Patients with a craniopharyngioma and an RCC were more likely to experience both transient and persistent DI. The increased risk of DI after resection of a craniopharyngioma has been well documented in the literature. 7,15,22 In the present series, the overall incidence of new persistent DI was 31%. This rate is relatively low considering that many surgeons believe the complication is unavoidable if complete resection is attempted; however, some authors 7 have found a lower-than-expected incidence if the operation is performed by a more experienced surgeon. Nevertheless, our data underscore the advantages of the transsphenoidal approach over the transcranial approach, in which the incidence of DI is much higher. 22 It has been previously postulated that a craniopharyngioma and an RCC may represent two points on a spectrum of sellar and parasellar cysts. 13,21 As such, it is not surprising that resection of these lesions is associated with an increased risk of both transient and persistent DI. In the present series, the observed rate of persistent DI after removal of an RCC was 9.7%, which is slightly higher than the rate in other series. 1,9 Note, however, that a rate comparison among the three series (9.7 compared with compared with 6.5% 1 ) is not statistically significant because of the small sample sizes and relatively low incidence of persistent DI. The surgical philosophy at the UVA Medical Center includes attempted radical resection as opposed to simple drainage for most cases. The advantage of this approach includes a decreased rate of recurrence. Nevertheless, any increased incidence of DI, should it become significant, may occur because of a difference in surgical philosophy. 21 Data from previous studies have shown an increased incidence of transient DI among patients with CD. 14 There are a number of possible explanations for this circumstance. First, CD is most often diagnosed through a variety of laboratory tests, including those indicating an elevated serum adrenocortical stimulating hormone level, elevated serum cortisol level with loss of diurnal variation, and increased urinary free cortisol level as well as the dexamethasonesuppression test. Often, patients with biochemical evidence of CD do not demonstrate a visible pituitary adenoma on formal intracranial imaging studies, including computerized tomography scanning or MR imaging. 24,25 Just as patients with a microadenoma may require more extensive exploration of the sella, which involves greater manipulation of the gland and stalk and an increased potential for damage, pa- 452

6 Diabetes insipidus and transsphenoidal surgery tients with CD and a hidden adenoma are equally at risk. This point has been made by others. 14,23,24 Indeed, the incidence of DI is greater in patients with CD and a microadenoma than in patients harboring a microadenoma alone (23.9 compared with 17.4%, respectively), although the study is not adequately powered for this difference to be statistically significant. Interestingly, the incidence of DI among patients with CD and a macroadenoma was greater than the overall incidence of DI among all patients with a macroadenoma alone. This circumstance indicates that other factors may also be at least partially responsible. Cortisol can induce a decrease in the renal concentrating ability, 14,20 and the consequent hypotonic polyuria may be misdiagnosed as DI. Additionally, impaired glucose tolerance or overt diabetes mellitus is commonly encountered in patients with CD. 31 Although meticulous glucose control was an essential part of the postoperative care in all patients, including those with CD, it is still possible in this retrospective study that glycosuria may occasionally have been misdiagnosed as DI. Conclusions Even with modern surgical techniques performed by an experienced surgeon, transient DI remains a common complication of transsphenoidal surgery. In contrast, persistent DI is relatively uncommon, especially in patients with pituitary adenomas. Patients with microadenomas and CD have an increased risk of suffering transient DI. This circumstance is likely caused by the more aggressive gland and stalk manipulations required during resection of the microadenoma as well as changes in sellar and parasellar anatomy in patients with macroadenomas. Patients with an observable intraoperative CSF leak have an increased risk of both transient and persistent DI. This increased incidence may simply be a marker of a more aggressive resection, although a direct inflammatory effect cannot be eliminated as a potential cause. Patients with a craniopharyngioma have an increased risk for both transient and persistent DI due to the invasive nature of these aggressive tumors and their intimate involvement with the pituitary stalk and the hypothalamus; however, our data demonstrate a lower-than-expected incidence, indicating that surgical experience may play a role in the avoidance of this complication. Rathke cleft cysts are also associated with an increased incidence of both transient and persistent DI. Acknowledgments We thank Ms. Rita Merkle and Ms. Lennie Clore for their secretarial assistance and Dr. Keith Littlewood for his helpful comments regarding the manuscript. References 1. Benveniste RJ, King WA, Walsh J, Lee JS, Naidich TP, Post KD: Surgery for Rathke cleft cysts: technical considerations and outcomes. J Neurosurg 101: , Black PM, Zervas NT, Candia GL: Incidence and management of complications of transsphenoidal operation for pituitary adenomas. Neurosurgery 20: , Camus J, Roussy G: Experimental researches on the pituitary body. Diabetes insipidus, glycosuria and those dystrophies considered as hypophyseal in origin. Endocrinology 4: , Ciric I, Ragin A, Baumgartner C, Pierce D: Complications of transsphenoidal surgery: results of a national survey, review of the literature, and personal experience. Neurosurgery 40: , Cohen C, Rice EN, Thomas DE, Carpenter TO: Diabetes insipidus as a hallmark neuroendocrine complication of neonatal meningitis. Curr Opin Pediatr 10: , Colombo N, Berry I, Kucharczyk J, Kucharczyk W, de Groot J, Larson T, et al: Posterior pituitary gland: appearance on MR images in normal and pathologic states. Radiology 165: , Duff JM, Meyer FB, Ilstrup DM, Laws ER Jr, Schleck CD, Scheithauer BW: Long-term outcomes for surgically resected craniopharyngiomas. Neurosurgery 46: , Elias WJ, Chadduck JB, Alden TD, Laws ER Jr: Frameless stereotaxy for transsphenoidal surgery. Neurosurgery 45: , el-mahdy W, Powell M: Transsphenoidal management of 28 symptomatic Rathke s cleft cysts, with special reference to visual and hormonal recovery. Neurosurgery 42:7 17, Fisher C, Ingram WR: The effect of interruption of the supraoptico-hypophyseal tracts on the antidiuretic, pressor and oxytocic activity of the posterior lobe of the hypophysis. Endocrinology 20: , Franco-Paredes C, Evans J, Jurado R: Diabetes insipidus due to Streptococcus pneumoniae meningitis. Arch Intern Med 161: , Gatell JM, Esmatjes E, Serra C, Aymerich M, SanMiguel JG: Diabetes insipidus and anterior pituitary dysfunction after staphylococcal meningitis and multiple brain abscesses. J Infect Dis 146: 102, Harrison MJ, Morgello S, Post KD: Epithelial cystic lesions of the sellar and parasellar region: a continuum of ectodermal derivatives? J Neurosurg 80: , Hensen J, Henig A, Fahlbusch R, Meyer M, Boehnert M, Buchfelder M: Prevalence, predictors and patterns of postoperative polyuria and hyponatraemia in the immediate course after transsphenoidal surgery for pituitary adenomas. Clin Endocrinol 50: , Honegger J, Buchfelder M, Fahlbusch R: Surgical treatment of craniopharyngiomas: endocrinological results. J Neurosurg 90: , Iraci G, Giordano R, Gerosa M, Pardatscher K, Tomazzoli L: Tuberculoma of the anterior optic pathways. Case report. J Neurosurg 52: , Jane JA Jr, Laws ER Jr: The surgical management of pituitary adenomas in a series of 3,093 patients. J Am Coll Surg 193: , Jane JA Jr, Thapar K, Kaptain GJ, Maartens N, Laws ER Jr: Pituitary surgery: transsphenoidal approach. Neurosurgery 51: , Kelly DF, Laws ER Jr, Fossett D: Delayed hyponatremia after transsphenoidal surgery for pituitary adenoma. Report of nine cases. J Neurosurg 83: , Kleeman CR, Levi J, Better O: Kidney and adrenocortical hormones. Nephron 15: , Laws ER, Kanter AS: Rathke cleft cysts. J Neurosurg 101: , 2004 (Editorial) 22. Lehrnbecher T, Muller-Scholden J, Danhauser-Leistner I, Sorensen N, von Stockhausen HB: Perioperative fluid and electrolyte management in children undergoing surgery for craniopharyngioma. A 10-year experience in a single institution. Childs Nerv Syst 14: , Olson BR, Gumowski J, Rubino D, Oldfield EH: Pathophysiology of hyponatremia after transsphenoidal pituitary surgery. J Neurosurg 87: ,

7 E. C. Nemergut, et al. 24. Olson BR, Rubino D, Gumowski J, Oldfield EH: Isolated hyponatremia after transsphenoidal pituitary surgery. J Clin Endocrinol Metab 80:85 91, Partington MD, Davis DH, Laws ER Jr, Scheithauer BW: Pituitary adenomas in childhood and adolescence. Results of transsphenoidal surgery. J Neurosurg 80: , Sane T, Rantakari K, Poranen A, Tahtela R, Valimaki M, Pelkonen R: Hyponatremia after transsphenoidal surgery for pituitary tumors. J Clin Endocrinol Metab 79: , Seckl J, Dunger D: Postoperative diabetes insipidus. Br Med J 298:2 3, Seckl JR, Dunger DB, Lightman SL: Neurohypophyseal peptide function during early postoperative diabetes insipidus. Brain 110: , Semple PL, Laws ER Jr: Complications in a contemporary series of patients who underwent transsphenoidal surgery for Cushing s disease. J Neurosurg 91: , Singer PA, Sevilla LJ: Postoperative endocrine management of pituitary tumors. Neurosurg Clin North Am 14: , Smith M, Hirsch NP: Pituitary disease and anaesthesia. Br J Anaesth 85:3 14, Wilson CB, Dempsey LC: Transsphenoidal microsurgical removal of 250 pituitary adenomas. J Neurosurg 48:13 22, 1978 Manuscript received December 8, Accepted in final form May 2, Address reprint requests to: Edward C. Nemergut, M.D., Department of Anesthesiology, University of Virginia Health System, P.O. Box , Charlottesville, Virginia en3x@ virginia.edu. 454