M. B. Linskey1, D. Brunke-Reese1, E. B. Lehman2, D. I. Soybel1, M. G. Kaag1,3 1Penn State University College Of Medicine,Department Of Surgery,Hershey, PA, USA 2Penn State University College Of Medicine,Department Of Public Health Sciences,Hershey, PA, USA 3Penn State University College Of Medicine,Division Of Urology,Hershey, PA, USA
Introduction: Post-operative hyperglycemia has been associated with adverse outcomes including increased length of stay (LOS) and increased costs of care. In the cardiac, vascular, general, and trauma surgery populations, post-operative hyperglycemia has also been linked to an increased risk of mortality. Patients without diabetes mellitus who develop acute hyperglycemia post-operatively are at an increased risk of complications compared to their counterparts with diabetes. Radical cystectomy for bladder cancer carries an inherent risk of post-operative morbidity due to the complexity of the procedure and the medical comorbidities of the patients. Morbidity of cystectomy includes frequent readmissions for renal failure, wound occurrences, ileus, failure to thrive, obstruction, and urinary tract infections. We investigate the impact of post-operative hyperglycemia on recovery following radical cystectomy (RC).
Methods: A retrospective chart review identified patients undergoing RC between May 2010 and December 2014 with at least one glucose level within 48 hours of surgery. Associations between post-operative hyperglycemia (defined as a first post-operative blood glucose >140mg/dL) and outcomes, including total hospital costs, LOS, and surgical site occurrences were determined.
Results: 176 patients underwent RC; 122 (69%) met our definition of post-operative hyperglycemia. 87 of 128 (68%) patients without diabetes, exhibited hyperglycemia postoperatively. 47 (54%) of these 87 patients required post-operative insulin, including 31 (36%) whose insulin requirement persisting beyond post-operative day 2. On univariate analysis, BMI classification predicted hyperglycemia (obese vs non-obese: Odds ratio (OR) 2.68, [95% Confidence Interval (CI) 1.25-5.75] p=0.01). This association was strong in patients without diabetes (OR 3.55 [95% CI 1.34-9.39] p=0.01), but not significant in those with diabetes. LOS (in days) was shorter in patients who were hyperglycemic post-RC regardless of prior diabetes diagnosis (Difference of medians (DOM) -2.0 [-3.5 to -0.5] p=0.01). This effect remained on multivariable analysis (DOM -2.19 [-3.54, -0.83] p=0.002) controlling for age, gender, race, Charlson score, ASA class, and BMI. Similarly, on multivariable analysis, hospital costs (in US dollars) were lower in patients with post-operative hyperglycemia (DOM -8,863.69 [-12,887.37, -4,840.17] p<0.001).
Conclusion: Post-operative hyperglycemia is common after RC and may occur in patients without diabetes. Contrary to results reported in the general surgery literature, hyperglycemia after RC was associated with shorter LOS and hospital costs. Whether this phenomenon is due to a protective effect associated with hyperglycemia, or is secondary to the aggressive post-operative management afforded these patients, is not yet clear.