In addition to the Pearl Diver-based retrospective study by Arshi et al. on one-year complications after outpatient knee replacement, the December 6, 2017 issue of JBJS contains a NSQIP-based retrospective study by Basques et al. that compares 30-day adverse events and readmissions among 1,236 patients who underwent same-day-discharge hip or knee (total or unicompartmental) arthroplasty with an equal number of propensity score-matched patients who were discharged at least 1 calendar day after the procedure.
When analyzing all three procedures together, the authors found no overall between-group differences in the rates of any adverse event (severe or minor) or readmission. However, when authors analyzed individual adverse events, the same-day group had decreased thromboembolic events and increased 30-day reoperations compared to inpatients. Analysis of individual procedures revealed an increased 30-day reoperation rate for same-day total knee arthroplasty (TKA), compared with inpatient TKA. Overall, infection was the most common reason for reoperation and readmission following same-day procedures.
As with the Arshi et al. study, the limitations of the database prevented these authors from accounting for physician or hospital volume. However, they did identify several preoperative patient characteristics that increased the risk of 30-day readmission among same-day patients, and from those findings Basques et al. concluded that “obese patients, older patients [≥85 years of age], and those with diabetes mellitus may not be appropriate candidates for same-day procedures.”
In the December 6, 2017 issue of The Journal, Arshi et al. report on a detailed analysis of a large administrative database, looking specifically at one-year complications associated with outpatient versus standard inpatient knee replacement. This type of analysis is crucial because of the rapidly growing interest in outpatient joint replacement among patients, payers, and the orthopaedic community.
The data convince me that these outpatient procedures should proceed, but with a little more caution. Although the absolute complication rates in both surgical settings were very low, after adjusting for age, sex, and comorbidities, the authors found a higher relative risk of several surgical and medical complications among outpatients—including component failure, infection, knee stiffness requiring manipulation under anesthesia, and deep vein thrombosis.
One important element that is lacking in this analysis is adjustment for surgeon/hospital volume. We know from important work by Katz and others that patients managed at centers and by surgeons with greater volumes of total knee replacement have lower risks of perioperative adverse events.
These results from Arshi et al. are definitely not a call to stop the expansion of outpatient joint replacement protocols. Instead, I think this study should prompt every joint-replacement center to analyze its risk-adjusted inpatient and outpatient outcomes—and to ensure, as these authors emphasize, that outpatients receive the same level of attention to rehabilitation, antibiotic administration, and thromboprophylaxis as inpatients.
Enhancing outpatient knee-replacement protocols will serve local communities well, and the nationwide orthopaedic community will receive further confirmation that outpatient joint replacement is a move in the right direction.
Marc Swiontkowski, MD
The practice of orthopaedic surgery is moving fairly rapidly to the outpatient environment. Advances in less invasive surgical procedures, regional anesthesia, and postoperative pain management have provided the foundation for this transition. The migration to outpatient surgery centers enables surgeons to use surgical teams more focused on orthopaedic technology and practice parameters. The concern that arises in everyone’s mind, though, is the issue of safety.
In the October 19, 2016 issue of JBJS, Qin et al. analyzed the NSQIP database and found that the outpatient surgical treatment of patients with a closed ankle fracture and minimal comorbidities resulted in lower risk of pneumonia and no difference in surgical morbidity, reoperations, and readmissions when compared with inpatient surgery.
The NSQIP dataset is voluntary and, as with any database, confounding variables are unavoidable. But these authors used propensity score matching and Bonferroni correction to minimize selection bias and manage multiple comparisons.
The study excluded emergency cases, cases with preoperative sepsis, and cases of open ankle fracture, and I can still foresee that patients with more severe fracture patterns, soft tissue compromise, and unstable medical comorbidities would be better off treated as inpatients. Nevertheless, it is reassuring that this study found no differences in complication or readmission rates. These findings reinforce the movement of orthopaedic surgical practice to the outpatient setting, and in my experience that movement is wholly welcomed by patients and their families.
Marc Swiontkowski, MD