SUN-461 Predictors of Early and Late Hyponatremia after Pituitary Surgery

Abstract

Abstract Introduction: Hyponatremia occurs in the first weeks after transsphenoidal surgery in 4-20% of patients and can lead to increased morbidity, hospital length of stay, and readmission rates. Age, sex, tumor size, and a diagnosis of Cushing disease have been variably suggested as predictive factors. Methods: We retrospectively reviewed sellar masses resected at our center from 2006-2012 prior to database lockdown to identify predictors of early (≤ 3 days) and late (≥ 4) postoperative hyponatremia. Numerical variables were summarized as mean ± SD and median (range), and significance testing calculated by Chi-square and t-test. Results: Of 294 sellar masses, 82% were pituitary adenomas (52% NFA, 16% GH, 10% PRL, 4% ACTH), 10% Rathke’s cleft cysts, and 8% nonpituitary masses. Hyponatremia developed early in 20% and late in 14%. In those with early hyponatremia, 15% occurred on the day of surgery, 53% on postoperative day 1, 20% on day 2, and 12% on day 3; median sodium was 132 mmol/L (116-134). Mean age was 54 years (SD 16) and 58% were female in those with early hyponatremia vs 51 years (SD 16) and 45% female in those without (RR 1.52, p=0.04). Median length of stay was 3.9 days (3-5) in those with early hyponatremia vs 3.9 (2-4) in those without (p=0.03). Mean maximal tumor diameter was 20 mm (SD 10) in both groups, although 62% of early hyponatremic patients had optic chiasm compression vs 48% without (RR 1.83, p=0.049). Tumor subtype did not differ between groups. Transient diabetes insipidus (DI) was seen in 20% with early hyponatremia vs 21% in those without; 33% of those with hyponatremia had lumbar drain placement vs 20% without (RR 1.66; p=0.03). In those with late hyponatremia, 25% occurred on postoperative days 4-6, 53% on day 7, and 22% on days 8-21; median sodium was 127 mmol/L (112-134). Mean age and sex distribution were similar with and without late hyponatremia. Late hyponatremia was seen in 22% with functioning adenomas vs 12% with NFAs; 30% with acromegaly had hyponatremia (RR 2.82; p=0.005). Subsequent hypopituitarism rates were similar, but inpatient desmopressin was administered in 41% of those with late hyponatremia vs 18% of those without (p=0.009). Readmission rate was 45% in those with late hyponatremia vs 13% in those without (p=0.0001). Use of antihypertensives, antiemetics, or opioids, and preoperative hypopituitarism and sodium levels were not predictive of early or late hyponatremia. Conclusions: Hyponatremia after pituitary surgery significantly prolongs hospital length of stay and results in high readmission rates. Those with optic chiasm compression and lumbar drain placement, especially women, are at increased risk of early hyponatremia, while acromegaly and transient post-operative DI are predictors of late hyponatremia. Closer monitoring and management of sodium levels and fluid restriction in patients with these risk factors may prevent readmission.