Delayed postoperative hyponatremia after transsphenoidal surgery: prevalence and associated factors

Transcription

1 J Neurosurg 119: , 2013 AANS, 2013 Delayed postoperative hyponatremia after transsphenoidal surgery: prevalence and associated factors Clinical article Namath S. Hussain, M.D., M.B.A., 1 Mackenzie Piper, B.A., 1 W. Grant Ludlam, B.A., 1 William H. Ludlam, M.D., Ph.D., 2 Cindy J. Fuller, Ph.D., 1 and Marc R. Mayberg, M.D. 1 1 Seattle Pituitary Center, Swedish Neuroscience Institute, Seattle, Washington; and 2 Novartis Pharmaceuticals Corporation, East Hanover, New Jersey Object. Transient delayed postoperative hyponatremia (DPH) after transsphenoidal surgery (TSS) is common and can have potentially devastating consequences. However, the true prevalence of transient symptomatic and asymp tomatic DPH has not been studied in a large patient cohort with close and accurate follow-up. Methods. A retrospective analysis of a single-institution prospective database was conducted; all patients undergoing TSS for lesions involving the pituitary gland were followed up in a multidisciplinary neuroendocrine clinic, and demographic, imaging, and clinical data were prospectively collected. Patients were examined preoperatively and followed up postoperatively in a standardized fashion, and their postoperative sodium levels were measured at Weeks 1 and 2 postoperatively. Levels of hyponatremia were rated as mild (serum sodium concentration meq/l), moderate ( meq/l), or severe (< 125 meq/l). Routine clinical questionnaires were administered at all postoperative office visits. Postoperative hyponatremia was analyzed for correlations with demographic and clinical features and with immediate postoperative physiological characteristics. Results. Over a 4-year interval, 373 procedures were performed in 339 patients who underwent TSS for sellar and parasellar lesions involving the pituitary gland. The mean (± SD) age of patients was 48 ± 18 years; 61.3% of the patients were female and 46.1% were obese (defined as a body mass index [BMI] 30). The overall prevalence of DPH within the first 30 days postoperatively was 15.0%; 7.2% of the patients had mild, 3.8% moderate, and 3.8% severe hyponatremia. The incidence of symptomatic hyponatremia requiring hospitalization was 6.4%. The Fisher exact test detected a statistically significant association of DPH with female sex (p = 0.027) and a low BMI (p = 0.001). Spearman rank correlation detected a statistically significant association between BMI and nadir serum sodium concentration (r = 0.158, p = 0.002) and an inverse association for age (r = , p = 0.031). Multivariate analyses revealed a positive correlation between postoperative hyponatremia and a low BMI and a trend toward association with age; there were no associations between other preoperative demographic or perioperative risk factors, including immediate postoperative alterations in serum sodium concentration. Patients were treated with standardized protocols for hyponatremia, and DPH was not associated with permanent morbidity or mortality. Conclusions. Delayed postoperative hyponatremia was a common result of TSS; a low BMI was the only clear predictor of which patients will develop DPH. Alterations in immediate postoperative sodium levels did not predict DPH. Therefore, an appropriate index of suspicion and close postoperative monitoring of serum sodium concentration should be maintained for these patients, and an appropriate treatment should be undertaken when hyponatremia is identified. ( Key Words transsphenoidal pituitary surgery hyponatremia diabetes insipidus Transsphenoidal surgery (TSS) for treating lesions of the anterior skull base is the mainstay surgical treatment for patients with sellar and parasellar lesions. However, TSS procedures are often complicated by electrolyte abnormalities in the perioperative period, including diabetes insipidus and hyponatremia. Diabetes insipidus Abbreviations used in this paper: ADH = antidiuretic hormone; BMI = body mass index; DPH = delayed postoperative hyponatremia; POD = postoperative day; SIADH = syndrome of inappropriate secretion of ADH; TSS = transsphenoidal surgery. J Neurosurg / Volume 119 / December 2013 results from a deficiency of antidiuretic hormone (ADH), presumably related to surgical stress, manipulation of the neurohypophysis, or both. 6,14,17 Hyponatremia after TSS usually results from hypocortisolemia or a syndrome of inappropriate secretion of ADH (SIADH). These conditions are related to altered free-water homeostasis, with resulting changes in the level of serum sodium. Serum sodium concentration and serum osmolality are maintained and modulated by several homeostatic mechanisms, including free-water intake by stimulated thirst, secretion of ADH, and renal filtering. 1,29 Sodium homeostasis, which 1453

2 N. S. Hussain et al. also affects the concentrations of other ions via the function of the sodium-potassium membrane ion pump, is vital to the normal physiological functioning of cells. Hyponatremia refers to a clinical condition in which serum sodium levels are below 135 meq/l. Management of hyponatremia is complex and requires specific knowledge of its etiology, and inadequate treatment may have serious consequences, including the development of central pontine myelinolysis, seizures, cardiac dysfunction, and death. 2,6,9,10,13,25,31,34 Severe hyponatremia is associated with significant morbidity and high mortality rates; a serum sodium level of less than 105 meq/l is associated with a mortality rate greater than 50%. 10 Symptomatic hyponatremia is relatively nonspecific in its presentation. Patients may report discomforts ranging from vague constitutional symptoms with nausea and vomiting to an altered level of consciousness. A correction of abnormal sodium levels that is too rapid can cause drastic shifts in intracerebral cell size, leading to permanent damage. 17,29 Thus, a methodical treatment plan is needed to determine whether the abnormality in sodium levels is due to excessive sodium excretion or SIADH; clinicians must accurately track postoperative sodium levels and treat patients quickly and appropriately. The phenomenon of delayed postoperative hyponatremia (DPH) after TSS has been described in several reports. 15,18,27,32,33 Many of these reports are limited retrospective studies based on patients readmitted with symptomatic hyponatremia. The prevalence of DPH varies among the reported case series, since symptoms of hyponatremia are often nonspecific, and patients discharged from the hospital may not contact their health care providers. In addition, the cohort size in these studies limits the ability to identify predictive demographic or perioperative factors associated with DPH. The goals of our study were to identify preoperative, intraoperative, and immediate postoperative characteristics that predict DPH after TSS. All surgical patients at a single multidisciplinary neuroendocrine clinic were treated according to a standardized regimen, and clinical data were prospectively collected. Patients were evaluated preoperatively and followed up postoperatively in a standardized fashion, and their postoperative sodium levels were routinely determined at 1 and 2 weeks postoperatively. Methods Patient Eligibility This study and the maintenance of a prospective database for pituitary surgery patients at Swedish Pituitary Center were approved by the Swedish Medical Center institutional review board and ethics committees. All patients undergoing TSS involving the pituitary gland for treatment of nonmalignant intrasellar and parasellar lesions at Swedish Neuroscience Institute from January 2007 to December 2010 were included in this study. Patients with lesions that were exclusively suprasellar or primarily arising from the clivus, planum sphenoidale, or cavernous sinus not involving the sella were not included in this study. Patients with a history of preoperative diabetes insipidus requiring medication were also excluded. Each surgery was treated as an independent event and repeat procedures were included in the analysis. A prospective database of all TSS patients included clinical and demographic indices, imaging characteristics, tumor histology, laboratory values, and postoperative complications. Standard Preoperative Evaluation All patients were evaluated and followed up in the Seattle Pituitary Center using a standardized protocol with uniform collection of their preoperative data. Laboratory assessments included standard pituitary hormone measurements and stimulation testing as indicated and metabolic, blood count, and coagulation studies. Preoperative imaging included thin-slice coronal and sagittal T1-weighted MRI of the sellar region with and without gadolinium and CT for image guidance. Anesthesia and Surgery All patients received a combination of remifentanil and isoflurane anesthesia. Perioperative supplementation of corticosteroids was administered only to patients with documented preoperatively impaired hypothalamicadrenal function. An endonasal-transseptal approach was performed in all cases using an operating microscope and selective endoscopy. Standard microsurgical techniques were used during tumor removal, and patients underwent placement of an autologous fat graft to prevent CSF leakage. Inpatient Perioperative Care After surgery, patients were transported to the ward and advanced to a regular diet. Vital signs, urine output, and urine and serum sodium levels and osmolality were monitored at the end of each nursing shift. Diabetes insipidus (defined as a urine output > 300 ml for 2 consecutive hours and a urine specific gravity < 1.002) was treated with ad lib oral fluid replacement or, if persistent, with intravenous or oral desmopressin. Patients with Cushing disease were monitored in the hospital for 72 hours after the surgery to document surgical remission; all other patients had planned discharges on postoperative day (POD) 1. Supplemental corticosteroids after surgery were administered only to those patients with preexisting hypocortisolism or to patients with a fasting serum cortisol of less than 15 μg/l on POD Outpatient Postoperative Care After discharge, all patients underwent at least 2 postoperative sodium determinations and were contacted via telephone by a neuroendocrine nurse practitioner on POD 6 and POD 13. Levels of hyponatremia were categorized as mild (serum sodium concentration of meq/l), moderate ( meq/l), or severe (< 125 meq/l). Management of hyponatremia was based on serum sodium levels in and degree of symptoms of the patients. Asymptomatic patients or symptomatic patients with mild hyponatremia underwent a repeat sodium de J Neurosurg / Volume 119 / December 2013

3 Postoperative hyponatremia after transsphenoidal surgery termination within 24 hours. Patients with symptomatic hyponatremia or moderate to severe hyponatremia were admitted to the hospital or evaluated in the clinic or emergency department. Outpatients with mild hyponatremia were managed as above with outpatient free-water restriction and sequential sodium determinations until normalization had occurred. Admitted patients underwent repeat fasting serum cortisol determination and simultaneous serum/urine electrolyte and osmolality determinations to confirm SIADH. To gradually normalize serum sodium levels (with a target increase in sodium < 8 10 meq/l per 24 hours), the patients were treated with monitored freewater restriction (1000 ml per 24 hours) and occasionally with judicious infusions of 2% NaCl or of vasopressin receptor antagonists (Vaprisol or tolvaptan). Patients were discharged when their serum sodium concentration was > 130 meq/l and symptoms were mild or had resolved. Statistical Analysis All statistical analyses were performed using SPSS The threshold for statistical significance was set a priori at a = Adjustments for multiple comparisons were made with the Bonferroni correction. Continuous data are presented as mean ± SD; categorical and nominal data are presented as frequency and percentage. Demographic and preoperative characteristics were recorded for individual patients, whereas perioperative and postoperative characteristics were determined for all procedures performed. Assumptions of normality were evaluated with the Kolmogorov-Smirnov test. Data were visually screened for outliers and departures from normality in frequency histograms and normal probability plots. Univariate ANOVA, chi-square or Fisher exact tests, and Student t-tests were used to interrogate individual characteristics that might be associated with postoperative hyponatremia; those that reached statistical significance were included in a general linear model for multivariate analysis. Results Characteristics of the Patient Cohort and Surgical Procedures From 2007 to 2010, 386 TSSs for nonmalignant lesions involving the pituitary gland were performed at the Swedish Neuroscience Institute by a single surgeon (M.R.M.). After excluding 13 patients with preoperative diabetes insipidus, 373 procedures in 339 patients (61.3% female and 38.7% male) were included in the analysis. The details of the patient demographics are provided in Table 1. Mean age at the time of surgery was 48 ± 18 years (range years); 46.1% of the patients were obese, 30.7% had a history of hypertension, 14.2% had diabetes mellitus, 28.0% had preoperative thyroid disease, and 3.2% had preoperative renal disease. Pathologies (Table 1) included nonfunctioning adenoma (113, 37.5%), adrenocorticotropin-secreting adenoma (104, 31.5%), prolactinoma (36, 12.0%), growth hormone secreting adenoma (29, 9.6%), Rathke cleft cyst (10, 3.3%), craniopharyngioma (7, 2.3%), and meningioma (4, 1.3%). J Neurosurg / Volume 119 / December 2013 Tumor size ranged from 0.1 to 4.5 cm in greatest diameter (mean 1.7 ± 1.0 cm). The average length of procedures was 61.8 ± 27.0 minutes (range minutes). Twentysix patients (7.7%) underwent repeat surgery within 30 days for a residual adrenocorticotropin-secreting tumor or CSF leakage. Postoperative Hyponatremia Postoperative laboratory data of physiological characteristics were available for 363 (97%) of the 373 procedures on POD 1, and outpatient laboratory determinations were available for 350 procedures (94%). The relative prevalence, severity, and timing of postoperative abnormalities in serum sodium concentration are shown in Fig. 1. The average serum concentration of sodium on POD 1 was 140 ± 4 meq/l (range meq/l). On POD 1, 7.8% of the patients showed symptoms of hyponatremia, 82.3% had normal sodium values, and 7.2% were hypernatremic (Na > 145 meq/l). Seventy-four patients (19.8%) had postoperative diabetes insipidus requiring 1 or more doses of desmopressin. Eighty-two patients (22.0%) had any episode of postoperative hyponatremia (either early or delayed). The overall prevalence of DPH was 15.0%; 7.2% of the patients had mild, 3.8% had moderate, and 3.8% had severe hyponatremia within 30 days (Fig. 1). The mean nadir serum sodium concentration within 30 days of TSS for the entire patient cohort was 138 meq/l (range meq/l); among patients with hyponatremia, the 30-day mean nadir sodium concentration was 128 meq/l (range meq/l). Hospitalization for hyponatremia was required after 24 procedures (6.4%). Factors Associated With DPH The results of univariate analysis of factors potentially associated with DPH are shown in Table 1. Delayed postoperative hyponatremia was not significantly associated with age, race, smoking history, alcohol use, hypertension, diabetes mellitus, preoperative hypopituitarism, thyroid disorders, or renal disease; an association was suggested between age and DPH (p = 0.088). Tumor type, repeat surgery, procedure duration, and serum sodium concentration on POD 1 were also not significantly associated with DPH. A lower body mass index (BMI) was a significant predictor of both immediate postoperative hyponatremia (p = 0.003) and any episode of delayed hyponatremia (p = 0.001). The univariate analysis also showed that females were more likely than males to experience DPH (p = 0.027). Only 22.9% of patients experiencing DPH were obese (that is, had a BMI > 30 kg/m 2 ); in contrast, DPH prevalence was 50.4% in nonobese patients. Spearman rank analysis showed that the BMI was significantly correlated with nadir serum sodium concentration (r = 0.158, p = 0.002). Age and nadir sodium concentration were also significantly correlated (r = , p = 0.031). Moreover, the size of the tumor and procedure duration were significantly and inversely correlated (r = , p = 0.014), and nonparametric analysis indicated that tumor size was significantly correlated with age (r = 0.361, p < 0.001). Nonobese patients were more likely to be rehospitalized for hyponatremia (10.2% vs 3.0%, p = 1455

4 N. S. Hussain et al. TABLE 1: Demographics and other characteristics of the patients in this study* Characteristic Hyponatremic Nonhyponatremic p Value no. of patients no. of procedures age (yrs) 51 ± ± female 38 (76.0) 170 (58.8) race Caucasian 43 (87.8) 248 (87.3) African American 0 (0.0) 5 (1.8) Hispanic 1 (2.0) 11 (3.9) Asian/Pacific Island 3 (6.1) 8 (2.8) Native American 0 (0.0) 3 (1.1) BMI in kg/m 2 (mean ± SD) 26.7 ± ± obesity 11 (22.9) 132 (50.4) current smoker 14 (28.0) 73 (26.4) current alcohol use** 29 (58.0) 136 (49.6) history of hypertension 14 (28.0) 90 (31.1) history of diabetes mellitus 9 (18.0) 38 (13.1) history of thyroid disease 12 (24.0) 83 (28.7) renal disease 1 (2.0) 10 (3.5) disease state nonfunctioning adenoma 21 (46.7) 92 (35.9) Cushing disease 12 (26.7) 70 (27.3) prolactinoma 3 (6.7) 33 (12.9) acromegaly 4 (8.9) 25 (9.8) Rathke cleft cyst 1 (2.2) 9 (3.5) preop hypopituitarism (any) 12 (21.4) 80 (25.2) tumor size in cm (mean ± SD) 1.6 ± ± repeat surgery (w/in 30 days) 6 (10.7) 20 (6.3) duration of op in mins (mean ± SD) 63 ± ± serum sodium on POD 1 (meq/l) a 139 ± ± * The mean age of all patients was 48 ± 18 years (range years). Data are shown as mean ± SD and relative frequency (%) unless indicated otherwise. Frequency data for demographics, patient histories, and disease characteristics were calculated on the basis of the number of patients (n = 339), whereas the frequency data for procedural and postoperative characteristics were calculated on the basis of the number of procedures (373) regardless of repeat surgeries. The p values were calculated using Student t-tests for continuous data or Fisher exact test for categorical data and were considered statistically significant at p < Race was recorded for 49 patients of the hyponatremic group and 284 patients of the nonhyponatremic group. The BMI was assessed in 48 and 262 patients of the hyponatremic and nonhyponatremic groups, respectively; the mean BMI of all patients was 30.6 kg/m 2 (range kg/m 2 ). Defined as a BMI of 30.0 kg/m 2. Smoking habit was recorded in all hyponatremic patients and in 276 of the nonhyponatremic patients. ** Current alcohol use was recorded in all hyponatremic patients and in 274 of the nonhyponatremic patients. The disease state was assessed in 45 and 256 patients of the hyponatremic and nonhyponatremic groups, respectively. Tumor size was assessed in 25 and 138 patients of the hyponatremic and nonhyponatremic groups, respectively; the mean tumor size in all patients was 1.7 cm (range cm). The number of repeat surgeries was recorded for all procedures in both groups. The duration was recorded for 48 and 301 patients of the hyponatremic and nonhyponatremic groups, respectively; the mean duration of the procedure in all patients was 62 minutes (range minutes). a The serum sodium concentration was determined in all patients of the hyponatremic group and 307 patients in the nonhyponatremic group; the mean serum sodium concentration on POD 1 in all patients was 140 meq/l (range meq/l) J Neurosurg / Volume 119 / December 2013

5 Postoperative hyponatremia after transsphenoidal surgery Fig. 1. Relative frequency (as percentage) of immediate (POD 1) and delayed (within 30 days) postoperative hyponatremia (rated as mild, moderate, or severe) in 339 patients (representing 373 procedures) undergoing TSS for lesions involving the pituitary gland ) and more likely to have moderate or severe hyponatremia (6.4% and 6.4% vs 4.2% and 1.8%, respectively, p = 0.02) than the obese patients. Univariate ANOVA was used to test the statistical significance of associations of a patient s sex, repeat surgery within 1 month, age, BMI, and length of surgery with nadir serum sodium concentration. The results of this analysis are shown in Table 2. In this analysis, BMI was significantly associated and age tended to be associated with nadir serum sodium concentration (p < and p = 0.061, respectively). A general linear model for multivariate analysis was used to further test these associations (Table 3), and this analysis showed that the BMI was significantly associated with nadir serum sodium concentration (p = 0.001), and age tended to be associated with nadir serum sodium concentration (p = 0.068). These results indicated that the BMI was the only variable consistently associated with nadir serum sodium concentration. Discussion Sodium plays an important role in cellular function and determining serum osmolality. Hyponatremia becomes symptomatic or clinically significant when it causes a drop in serum osmolality. 5 Hypo-osmolality (a serum osmolality < 280 mosm/kg) signals excess total body water relative to solutes in the extracellular fluid, which may be due to solute depletion, dilution, or both. 4 Normally, the kidneys excrete up to L of free water per day, and osmoregulation mechanisms of the brain detect and adapt to hyponatremia with decreased thirst within minutes, and this adaptation is maximal by 2 3 days. 29 This inhibits the secretion of ADH from the neurons in the hypothalamus that constantly sense serum osmolality, leading to increased elimination of water in the urine. Clinicians must also keep in mind other causes of hyponatremia that must be addressed, including adrenal insufficiency, alcoholism, cirrhosis, hypothyroidism, and cardiogenic pulmonary edema. 4 Disorders of serum electrolytes and serum osmolality after TSS have been attributed to manipulation or injury of the neurohypophysis producing a triphasic response characterized by diabetes insipidus followed by SIADH, with either resolution or recurrent diabetes insipidus. 12,19 Ultmann et al. produced radiofrequency-induced partial injuries to the neurohypophysis in rats and noted delayed hyponatremia in 20 of 35 animals, 12 of which progressed to diabetes insipidus. 28 The hyponatremia was related to increased serum vasopressin levels consistent with SIADH. Clinical trials show, however, that increased vasopressin secretion is uncommon in hyponatremic patients after TSS. 17,27,31 We did not observe any association between postoperative diabetes insipidus and the development of subsequent hyponatremia, and only 1 previous study has shown such a relationship. 33 Olson and colleagues, in a study of 92 patients, prospectively monitored postoperative water balance and sodium dysregulation after pituitary surgery. 21 The authors observed that after a water-loading challenge, only a third of the hyponatremic patients had normal vasopressin suppression, and that other factors, such as changes in natriuresis and lower dietary sodium intake, contributed to hyponatremia in these patients. Previous studies have reported a varying incidence of DPH after TSS (Table 4). Many of these studies were limited by a small cohort size, nonuniform monitoring of postoperative sodium levels, and lack of correlation with factors potentially associated with the development of hyponatremia. Our aim was to collect prospective data in a standard manner to facilitate a more comprehensive analysis of the true prevalence of DPH. A rigorous follow-up protocol with postoperative sodium checks at 1 and 2 weeks postoperatively was designed to ensure the inclusion of asymptomatic or unrecognized cases of DPH that may have gone undetected in previous screening protocols. Laboratory values were available in a significant proportion of the cohort (97% on POD 1 and 94% of the outpatient assessments), reducing the likelihood of sampling bias in these findings. The overall prevalence of hyponatremia in our study was 22% (82 occurrences for 373 procedures), of which 15% occurred after discharge from the hospital; hospitalization was required after 6.4% of the procedures. Many of the DPH patients were asymptomatic or mildly symptomatic at the time of diagnosis, TABLE 2: Univariate ANOVA of nadir serum sodium concentration within 30 days of TSS* Characteristic F Value p Value Observed Power age BMI < * Between-case factors were sex of the patient and repeat surgery within 1 month. Continuous variables included in the model were age, BMI, and duration of surgery. A p < 0.05 was considered statistically significant. J Neurosurg / Volume 119 / December

6 TABLE 3: General linear model multivariate analysis of nadir serum sodium concentration within 30 days of TSS* Characteristic F Value p Value Observed Power age BMI N. S. Hussain et al. * Between-case factors were sex of the patient and repeat surgery within 1 month. Continuous variables included in the model were age, BMI, and duration of surgery. A p < 0.05 was considered statistically significant. and often these patients attributed any symptoms to other causes. In a retrospective analysis of 2297 patients who underwent TSS, Taylor et al. noted 53 patients (2.3%) who were treated for symptomatic hyponatremia. 27 The authors surmised that the incidence of hyponatremia was probably underestimated in their study owing to incomplete identification of affected patients. Our results indicated that hyponatremia is a much more common phenomenon after surgery than previously thought, and that many cases of hyponatremia might be missed by less comprehensive screening protocols. A comparison of previous studies reporting prevalence of DPH after TSS is shown in Table 4. The reported prevalence of DPH ranged from 2.3% to 53%; 15,18,27,32,33 most of these analyses were retrospective, and clearly the level of suspicion and thoroughness of the monitoring protocol were determining factors in the identification of DPH. In a prospective study of cases with nonfunctioning pituitary adenomas, Chen et al. 7 observed postoperative hyponatremia in 22.1% of these cases, comparable to the results of our study. Using preoperative sodium levels as a baseline, postoperative declines in serum sodium levels were observed in 92.0% of TSS patients. 16 The timing of the DPH after TSS has been generally consistent, occurring on PODs 4 7 and resolving within 2 3 weeks. 15,18,20,27,30,32,33 Factors associated with the development of DPH vary in previous reports and include age, 16,18,30 female sex 33 or estrogen use, 22 tumor size, 15,26,30 early diabetes insipidus after surgery, 33 Cushing disease, 14,24 and surgical trauma to the neurohypophysis. 22 Some of these factors were also identified in our univariate analyses, including female sex (p = 0.027). However, the multivariate analysis did not detect a significant association of DPH with a patient s sex, and indicated a suggestive but nonsignificant association of DPH with age (p = 0.068). Repeat surgery was not significantly associated with DPH in the multivariate and univariate analyses. Only a low BMI was robustly associated with the development of DPH (p = 0.001). The physiological basis for the association between low BMI and DPH remains unknown. Neuroendocrine factors commonly associated with BMI, such as Cushing disease and diabetes mellitus, were excluded in the current analysis. There is a strong association between obesity and impaired renal function; 3 patients with a higher BMI may be less responsive to elevated vasopressin levels after TSS, and thus be less likely to develop DPH. Obesity is associated with abnormal vasopressin secretion 8 and elevated copeptin (a fragment of the vasopressin prohormone), 11 further suggesting a relationship between BMI and a dysregulated vasopressin system. Treatment of DPH in the current series consisted of early prescription and implementation of free-water restriction for most patients with hyponatremia. Intravenous hypertonic saline infusion or vasopressin antagonists were reserved for a small number of patients with severe, refractory hyponatremia. We limited the use of more aggressive measures to correct serum sodium concentration, since rapid overcorrection of hyponatremia at rates greater than 10 meq/l per 24 hours has been associated with the development of central pontine myelinolysis. 9,23 TABLE 4: Overview of published case series of DPH after TSS* Authors & Year No. of Cases Delayed Hyponatremia (no. [%]) Associated Factors Lee et al., (18.0) age 50 yrs Taylor et al., (2.3) none Wei et al., (38.8) age, tumor size Chen et al., (22.1) none Zada et al., (23.0) female patients, previous transient DI Staiger et al., (53.0) tumor size Kelly et al., (9.0) tumor size Olson et al., (21.0) posterior pituitary surgical trauma, estrogen use Kinoshita et al., (30.7) age >60 yrs Kristof et al., (36.7) none Adams et al., (8.8) none Hensen et al., (2.4) Cushing disease Sane et al., (35.0) Cushing disease * DI = diabetes insipidus J Neurosurg / Volume 119 / December 2013

7 Postoperative hyponatremia after transsphenoidal surgery In the present study, there was no long-term morbidity associated with DPH using the conservative treatment measures described above. The present analysis describes the prevalence of DPH and associated risk factors from a large cohort of patients with carefully monitored serum sodium levels after TSS. However, there were several limitations to our study. Although the data collection was prospective, the retrospective analysis and single-institution design introduced a potential bias and lack of generalization to broader cohorts. Also, serum sodium levels were measured at discrete time points rather than continuously. Thus, we may have missed other time points when patients were hyponatremic. Finally, there may have been other variables not collected in this database that may have contributed to the development of DPH. Conclusions Hyponatremia continues to be a common complication associated with TSS involving the pituitary gland. The prevalence of DPH after TSS was higher than previously thought. Thus, a higher index of suspicion must be maintained for this clinical entity, and it is vitally important to monitor TSS patients after surgery because of the serious consequences of untreated or undertreated hyponatremia. Protocols to monitor serum sodium levels after TSS should be implemented in this patient population. Attention must be paid to a low preoperative BMI and to age, as these 2 variables were significantly or tended to be associated with DPH. We recommend that serum sodium concentration be measured in all patients in the immediate postoperative period and at least once during the period 5 10 days after surgery. Patients should be educated about the signs and symptoms of hyponatremia and additional evaluations should be performed when routine screenings or symptoms indicate hyponatremia. Acknowledgments We would like to acknowledge the assistance of Chris Kaperak in data collection and Mary Caverly and Karen Pabillon in manuscript preparation. Disclosure The authors report no conflict of interest concerning the materials or methods used in the study or the findings specified in this paper. During the study period, Dr. W. H. Ludlam was Director of the Neuroendocrine Service at Swedish Neuroscience Institute. His current position at Novartis Pharmaceuticals is unrelated to the topics investigated in this manuscript. Author contributions to the study and manuscript preparation include the following. Conception and design: Mayberg, WG Ludlam, WH Ludlam. Acquisition of data: all authors. Analysis and interpretation of data: Mayberg, Hussain, Fuller. Drafting the article: all authors. Critically revising the article: all authors. Reviewed submitted version of manuscript: all authors. Approved the final version of the manuscript on behalf of all authors: Mayberg. Statistical analysis: Fuller. References J Neurosurg / Volume 119 / December Adams JR, Blevins LS Jr, Allen GS, Verity DK, Devin JK: Disorders of water metabolism following transsphenoidal pituitary surgery: a single institution s experience. Pituitary 9: 93 99, Albanese A, Hindmarsh P, Stanhope R: Management of hyponatraemia in patients with acute cerebral insults. Arch Dis Child 85: , Amann K, Benz K: Structural renal changes in obesity and diabetes. Semin Nephrol 33:23 33, Atkin SL, Coady AM, White MC, Mathew B: Hyponatraemia secondary to cerebral salt wasting syndrome following routine pituitary surgery. Eur J Endocrinol 135: , Berker M, Hazer DB, Yücel T, Gürlek A, Cila A, Aldur M, et al: Complications of endoscopic surgery of the pituitary adenomas: analysis of 570 patients and review of the literature. Pituitary 15: , Boehnert M, Hensen J, Henig A, Fahlbusch R, Gross P, Buchfelder M: Severe hyponatremia after transsphenoidal surgery for pituitary adenomas. Kidney Int Suppl 64:S12 S14, Chen L, White WL, Spetzler RF, Xu B: A prospective study of nonfunctioning pituitary adenomas: presentation, management, and clinical outcome. J Neurooncol 102: , Coiro V, Chiodera P: Effect of obesity and weight loss on the arginine vasopressin response to insulin-induced hypoglycaemia. Clin Endocrinol (Oxf) 27: , Decaux G, Soupart A: Treatment of symptomatic hyponatremia. Am J Med Sci 326:25 30, Dubois GD, Arieff AI: Treatment of hyponatremia: the case for rapid correction, in Narins RG (ed): Controversies in Nephrology and Hypertension. New York: Churchill Livingstone, 1984, pp Enhörning S, Bankir L, Bouby N, Struck J, Hedblad B, Persson M, et al: Copeptin, a marker of vasopressin, in abdominal obesity, diabetes and microalbuminuria: the prospective Malmö Diet and Cancer Study cardiovascular cohort. Int J Obes (Lond) 37: , Finken MJ, Zwaveling-Soonawala N, Walenkamp MJ, Vulsma T, van Trotsenburg AS, Rotteveel J: Frequent occurrence of the triphasic response (diabetes insipidus/hyponatremia/diabetes insipidus) after surgery for craniopharyngioma in childhood. Horm Res Paediatr 76:22 26, Goldberg A, Hammerman H, Petcherski S, Zdorovyak A, Yalonetsky S, Kapeliovich M, et al: Prognostic importance of hyponatremia in acute ST-elevation myocardial infarction. Am J Med 117: , Hensen J, Henig A, Fahlbusch R, Meyer M, Boehnert M, Buch felder M: Prevalence, predictors and patterns of postoperative polyuria and hyponatraemia in the immediate course after transsphenoidal surgery for pituitary adenomas. Clin Endocrinol (Oxf) 50: , Kelly DF, Laws ER Jr, Fossett D: Delayed hyponatremia after transsphenoidal surgery for pituitary adenoma. Report of nine cases. J Neurosurg 83: , Kinoshita Y, Tominaga A, Arita K, Sugiyama K, Hanaya R, Hama S, et al: Post-operative hyponatremia in patients with pituitary adenoma: post-operative management with a uniform treatment protocol. Endocr J 58: , Kristof RA, Rother M, Neuloh G, Klingmüller D: Incidence, clinical manifestations, and course of water and electrolyte metabolism disturbances following transsphenoidal pituitary adenoma surgery: a prospective observational study. Clinical article. J Neurosurg 111: , Lee JI, Cho WH, Choi BK, Cha SH, Song GS, Choi CH: Delayed hyponatremia following transsphenoidal surgery for pituitary adenoma. Neurol Med Chir (Tokyo) 48: , Mamelak AN, Carmichael J, Bonert VH, Cooper O, Melmed S: Single-surgeon fully endoscopic endonasal transsphenoidal surgery: outcomes in three-hundred consecutive cases. Pituitary 16: ,

8 N. S. Hussain et al. 20. Marko NF, Weil RJ: A comparative effectiveness analysis of alternative strategies to assess hypothalamic-pituitary-adrenal axis function after microsurgical resection of pituitary tumors. Neurosurgery 68: , Olson BR, Gumowski J, Rubino D, Oldfield EH: Pathophysiology of hyponatremia after transsphenoidal pituitary surgery. J Neurosurg 87: , Olson BR, Rubino D, Gumowski J, Oldfield EH: Isolated hyponatremia after transsphenoidal pituitary surgery. J Clin Endocrinol Metab 80:85 91, Rahman M, Friedman WA: Hyponatremia in neurosurgical patients: clinical guidelines development. Neurosurgery 65: , Sane T, Rantakari K, Poranen A, Tähtelä R, Välimäki M, Pelkonen R: Hyponatremia after transsphenoidal surgery for pituitary tumors. J Clin Endocrinol Metab 79: , Semple PL, Laws ER Jr: Complications in a contemporary series of patients who underwent transsphenoidal surgery for Cushing s disease. J Neurosurg 91: , Staiger RD, Sarnthein J, Wiesli P, Schmid C, Bernays RL: Prognostic factors for impaired plasma sodium homeostasis after transsphenoidal surgery. Br J Neurosurg 27:63 68, Taylor SL, Tyrrell JB, Wilson CB: Delayed onset of hyponatremia after transsphenoidal surgery for pituitary adenomas. Neurosurgery 37: , Ultmann MC, Hoffman GE, Nelson PB, Robinson AG: Transient hyponatremia after damage to the neurohypophyseal tracts. Neuroendocrinology 56: , Vaidya C, Ho W, Freda BJ: Management of hyponatremia: providing treatment and avoiding harm. Cleve Clin J Med 77: , Wei T, Zuyuan R, Changbao S, Renzhi W, Yi Y, Wenbin M: Hyponatremia after transspheniodal surgery of pituitary adenoma. Chin Med Sci J 18: , Whitaker SJ, Meanock CI, Turner GF, Smythe PJ, Pickard JD, Noble AR, et al: Fluid balance and secretion of antidiuretic hormone following transsphenoidal pituitary surgery. A preliminary series. J Neurosurg 63: , Wise BL: Delayed hyponatremia after transsphenoidal surgery. J Neurosurg 85:991, 1996 (Letter) 33. Zada G, Liu CY, Fishback D, Singer PA, Weiss MH: Recognition and management of delayed hyponatremia following transsphenoidal pituitary surgery. J Neurosurg 106:66 71, Zafonte RD, Mann NR: Cerebral salt wasting syndrome in brain injury patients: a potential cause of hyponatremia. Arch Phys Med Rehabil 78: , 1997 Manuscript submitted March 5, Accepted August 6, Please include this information when citing this paper: published online September 20, 2013; DOI: / JNS Address correspondence to: Marc R. Mayberg, M.D., Swedish Neuroscience Institute, th Ave., Ste. 500, Seattle, WA marc.mayberg@swedish.org J Neurosurg / Volume 119 / December 2013