Predictors of Mortality and Morbidity in Patients undergoing Transsphenoidal Surgery for Treatment of Acromegaly

Introduction: Several studies have shown that patient predictors including race, payer status, socioeconomic class, hospital volume, and surgeon experience are key predictors of postoperative complications, mortality, length of stay, and hospital charges. To date, no study has used the Nationwide Inpatient Sample (NIS) data to selectively study predictors of outcome and disparities in patients seeking treatment of acromegaly from growth hormone producing pituitary adenomas.

Methods: This study utilized longitudinal hospital inpatient discharge data from 2002 to 2010 in the Healthcare Cost and Utilization Project NIS database. Acromegalic patients following transsphenoidal procedures were analyzed for endocrine complications (diabetes insipidus, panhypopituitarism, and electrolyte abnormalities), nonendocrine complications (intracerebral hemorrhage or hematoma, cranial nerve palsy, cerebrospinal fluid rhinorrhea, cerebral arteriogram, mechanical ventilation, blood transfusion, and deep venous thrombosis/pulmonary embolism/inferior vena cava filter installation). Discrete patient predictors, including race, payer status, and gender were encoded as categorical variables. Annual transsphenoidal procedure volume was calculated and reported in categorical tertiles (low,< 2; medium, 2–5; high,> 5 procedures) for analysis.

Results: From 2002 to 2010, there were 4,696 transsphenoidal surgeries involving 4,659 acromegaly cases available for the analysis. Of these, 1,090 (23.4%) had at least one endocrine or nonendocrine complication. The 4,696 procedures (47 biopsies, 3,492 partial resections, and 1,157 total resections) resulted in 610 discharges (12.99%) with an endocrine complication and 495 discharges (10.54%) with nonendocrine complications. There were no significant relationships between race, hospital volume, and other variables in predicting in hospital mortality. The majority of patients were white and privately insured. There were no significant relationships identified in predicting routine versus nonroutine hospital discharges. Most procedures were performed at high procedural volume urban teaching hospitals. In terms of endocrine complications, African American patients had an increased risk of diabetes insipidus (RR = 3.31; 95% CI = 1.44, 7.63; p value < 0.01) as compared with white patients. Patients seen at high volume centers had a decreased risk of postoperative panhypopituitarism (RR = 0.07; 95% CI = 0.07, 0.07; p value < 0.01). Hispanic patients also had an increased risk of nonendocrine complications (RR = 2.26; 95% CI = 1.34, 3.81; p value < 0.01). African American and Hispanic patients had increased hospital charges (RR, 2.83; 95% CI = 1.14, 6.99; p value = 0.02) and (RR = 2.11; 95 % CI= 1.24, 3.61; p value < 0.01), respectively, as compared with white patients. Hispanic patients had a median length of stay of 4.61 days which was significantly increased as compared with white patients (RR = 1.52; 95% CI = 1.09, 2.10; p value = 0.01).

Conclusion: NIS data show that race is a significant predictor of endocrine complications, nonendocrine complications, increased hospital charges, and prolonged length of stay in acromegaly patients undergoing transsphenoidal surgery. Hispanic and African-American patients with acromegaly may face ongoing disparities and barriers to accessing quality treatment for acromegaly.