Background: Patients with End-stage renal disease (ESRD) who are dialysis-dependent have an elevated risk of peri-operative morbidity and mortality. There are many scoring tools that can help predict perioperative risk. Unfortunately, none were validated specifically for patients with ESRD. We aimed to compare the performance of the Revised cardiac risk (RCRI) and the AUB-HAS2 scores in this specific patient population.
Methods: We performed a retrospective study of patients that had undergone surgery and entered the Acute Care Surgery National Surgical Quality Improvement Program (NSQIP) database from 2008 until 2012. We only included those years due to specific elements needed for score calculations. We included only patients with ESRD prior to surgery. Since the RCRI would include creatinine measurement, we did not include that variable. We analyzed the associations of the different scores with the outcomes of interest in the composite: Death, Myocardial infarction, or Stroke.
Results: There were 1,167,278 surgical procedures entered in the NSQIP database from 2008 until 2012. Of those, 32,337 patients had ESRD before surgery and therefore were selected for analysis. The age was 61.1 years (14.5), the BMI was 28.7 (8) and there were 13,958 females (43%). 15,331 patients were diabetic (47%), 27,506 had hypertension (85%) and 5,987 were current smokers (18.5%). 1,055 patients reported a history of Myocardial infarction in the past (3.3%), 2,199 had congestive heart failure (7%) 3,661 had a Percutaneous cardiac intervention previously (15%) and 3,680 (11.4%) had a history of stroke. Prior to admission, 6,366 patients reported angina symptoms or dyspnea (20.2%). The most common surgical category was General surgery (15,193 patients, 47.5%) followed by vascular surgery (13,856 (43.4%) then orthopedic surgery (1,164 patients 3.6%). 5,675 patients underwent emergency surgery (17.7%).3,190 patients died during hospitalization (10%), and 6,996 (21.6%) had the composite outcome (Death, MI, or stroke). For RCRI there were 8,664 (27%) with scores of 0, 12,769 (39.5%) with 1, 8,367 (25.9%) with 2, and 2,537 (7.8%) with 3. Mortality increased with higher RCRI: in group 0 it was 4.4%, in group 1 was 9.9%, in group 2: it was 12.8%, and in group 3:19.2%. The composite Endpoint also increased from 4.4% to 15.5% to 32.3% to 77% respectively. For the AUB-HAS-2, there were 2,430 with scores of 0 (7.5%), 9,192 (28.4%) with 1, 10,380 (32.1%) with 2, 6981(21.6%) with 3, and 3,3354 with greater than 3. Mortality was 1.7%, 4.1%, 9.66%, 14.6%, and 23. % in groups respectively. The composite endpoint was 8%, 11.7%, 20%, 31.5%, and 43.2% respectively. When we compared the performance of the RCRI and the AUB-HAS using an Area under the curve for mortality, AUBHAS performed better: 0.68[0.67-0.69] vs. RCRU 0.62[0.61-0.63] (p< 0.001), but for the composite endpoint, RCRI performed better 0.77 [0.765-0.777) vs. 0.677[0.663-0.677] (p< 0.001). This difference in the score's performance was maintained in most surgical category subgroups (General, Vascular, Orthopedic), and type of surgery (Cholecystectomy, Colon surgery, Breast surgery).
Conclusions: Patients with ESRD undergoing surgery are very likely to have poor outcomes. The RCRI (without Creatinine) and AUB HAS2 scores can provide guidance and estimate the AUB-HAS2 performed better for the outcome mortality and the RCRI was best for the composite outcome (Death, Myocardial infarction, and stroke).