Assessing perioperative risk is a common challenge in daily internal medicine practice. Risk models are sparse and focus primarily on cardiac and pulmonary risk. Little has been established regarding the implications of more common diagnoses. Affecting 23.5 million Americans, diabetes mellitus has long been associated with poor wound healing. In the American College of Surgeon's National Surgical Quality Improvement project (ACS‐NSQIP) a simplified data point regarding diabetes is gathered, splitting diabetics into 3 categories: none, oral diabetic treatment, and insulin. We hypothesized that diabetes would represent an important diagnosis in stratifying risk.
We utilized the ACS‐NSQIP public use files (PUF) from 2005 to 2008. This was done under their data use agreement and under the supervision of our institutional review board. Patient with diabetes were compared with those without including demographics and outcomes. Data were evaluated using chi‐square analysis in SPSS. Unless otherwise specified, significance of all reported proportions and odds ratios was P < 0.001.
Of 635,265 patients in the data set, 85.5% did not have diabetes (NoDM), 8.7% took oral medication (ORAL), and another 5.8% required insulin (IRDM). African Americans comprised 9.8% of the study population. Among African Americans in the data set, DM was more prevalent (OR, 1.7) and there was a far greater rate of IRDM (OR, 2.15) than in the remaining population. Patients older than 60 years had higher rates of diabetes (OR, 2.31) and IRDM (1.89) than those younger than 60. Men had higher rates of IRDM (6.5%) than women (5.2%). DM patients were more likely to experience postoperative complications (OR, 1.88) and mortality (OR, 2.15) than NoDM patients. In addition, the subset of IRDM patients experienced significantly higher rates of complications (OR, 2.37) and mortality (OR, 2.87). IRDM patients experienced more wound occurrence (9.9%), followed by ORAL patients (6.7%) and then NoDM patients (4.9%); P < 0.001. Patients with IRDM also experienced more infections (13.3%) than ORAL patients (9.5%) and NoDM patients (6.9%). The overall mortality rate was 1.7%. The rate was 3.2% in diabetics, far higher in IRDM patients, at 4.4%, compared with 1.7% in ORAL patients and 1.5% in NoDM patients. Examining the inpatient subset of the population revealed similar results. A total of 377,594 patients were analyzed, of whom 17.6% were diabetic and 7.5% were IRDM patients. Diabetics again displayed higher mortality (OR, 1.73) and postoperative complications (OR, 1.52), with IRDM patients typically showing an even higher likelihood of mortality (OR, 2.20) and complications (OR, 1.83).
Diabetes is predicted to affect half of all Americans by 2020. Even gross classification of this common disease can result in significant predictive value and aid in the process of assessing operative risk and counseling patients. Further work is warranted to develop simplified but broader risk models to predict and modify perioperative risk in these patients.
M. Farooq ‐ none; P. Watson ‐ none; I. Rubinfeld ‐ none
To cite this abstract:Watson P, Farooq M, Rubinfeld I. Diabetes and Perioperative Risk Evaluation in Medical Practice. Abstract published at Hospital Medicine 2011, May 10-13, Dallas, Texas. Abstract 138. Journal of Hospital Medicine. 2011; 6 (suppl 2). https://www.shmabstracts.com/abstract/diabetes-and-perioperative-risk-evaluation-in-medical-practice/. Accessed March 31, 2020.