Abstract
Introduction
Rates of readmission and reoperation following primary total knee arthroplasty (TKA) are under scrutiny due to new payment models, which penalize these negative outcomes. Many prior studies have demonstrated the risk of perioperative complications among patients with obesity. However, an elevated complication rate among patients with poor nutrition, as measured by hypoalbuminemia, has also been reported. We sought to determine whether controlling for hypoalbuminemia would mitigate the effect that prior authors had identified in patients with obesity. In addition, we hoped to identify an albumin threshold above which risk of readmission and reoperation would be minimized.
Materials and Methods
We reviewed the American College of Surgeons National Surgical Quality Improvement Program (ACS-NSQIP) database for the period of 2008–2016 to evaluate the rates of perioperative complications among primary TKA patients by obesity category and albumin level. Patients were divided into groups according to World Health Organization obesity category and continuous as well as dichotomous albumin level. The primary outcomes were readmission and reoperation within 30 days of surgery. Patients without BMI or albumin data were excluded, as were patients undergoing surgery for the treatment of fracture or bone tumor. Univariate logistic regression compared predictive value of body mass index and albumin on incidence of these outcomes, using the normal weight and albumin > 3.5 g/dL groups as controls. Multivariate logistic regression modeling controlled for age, gender, and comorbidity status. Receiver operator curves (ROC) were generated to determine if an albumin threshold could be determined, above which risk for these complications would be minimized.
Results
Outcome data was available 108,601 patients. Overall, 3,508 patients (3.46%) were readmitted and 1,219 patients (1.19%) underwent reoperation within 30 days. Hypoalbuminemia was present in 4,327 patients (4.11%), lowest in the Overweight group (3.16%) and highest in the Obese Type III (6.72%) and Underweight (12.62%) groups (p <0.001). Readmission and reoperation rates were highest among patients categorized as Obesity Class III (4.15 and 1.73%, respectively) compared with Normal Weight (3.59 and 1.14%, respectively, p <0.001). Readmission and reoperation rates were higher among patients with hypoalbuminemia (6.10 and 1.96%, respectively) compared with normal albumin (3.34 and 1.16%, respectively, p <0.001). After controlling for covariates of interest using multivariate regression including continuous albumin, odds ratio of readmission among the highest obesity category no longer differed from the Normal Weight control group. Odds ratio of reoperation among the highest obesity category declined from 1.52 to 1.38 after controlling for albumin and other covariates (p = 0.022). A subanalysis was performed excluding all patients with albumin < 3.5 g/dL, but results did not differ significantly from the multivariate model. ROC modeling was unable to identify a threshold for continuous BMI or albumin that would adequately mitigate risk for readmission or reoperation.
Discussion
Albumin and obesity remain significant independent risk factors for reoperation following primary TKA procedures. However, after controlling for hypoalbuminemia and other important covariates, readmission rates did not significantly differ between patients with Normal Weight and any other obesity category. Controlling for albumin alone does not mitigate all risk imposed by obesity on perioperative outcomes. Future studies may assess whether preoperative albumin repletion or weight loss may improve outcomes for patients with hypoalbuminemia and obesity.
For figures, tables, or references, please contact authors directly.
