(P-395) Evaluation of Drain Use During Total Knee Replacement Associated with Health Outcomes and Healthcare Utilization

Background: Drains are sometimes placed during knee replacement procedures to offset risk of hematomas and infection during wound healing, among other outcomes.

Objectives: Understanding which patients receive drains and which ones do not appears limited. Research on drain use may be useful in understanding their role on patient health outcomes, including factors associated with cost such as procedure time and potential for penalties incurred due to readmission.

Methods: Data from the Mercy Health electronic health records database from 2010-2020 was extracted for analysis. Patients aged 18 and older who underwent a primary total knee replacement for the treatment of osteoarthritis were eligible for inclusion. Additional exclusions were applied. Assessment of prevalence of outcomes as related to cemented and cementless designs were quantified using descriptive statistics.

Results: A total of 31,352 records were included, of which 4,222 had drain use (1,731 were cemented and 15 were cementless, 2,476 were unknown) and 27,130 had no-drain use (12,708 were cemented and 444 were cementless, 13,978 were unknown). On average, the sample was aged 67 years (SD=10) for both the drain and no-drain use groups. Provider volume, defined as number of TKA procedures the provider has performed in a year prior to index TKA, was generally higher in the no drain group (Mean=106, SD=68) compared to the drain group (Mean=77 procedures, SD=56). General infection at 1-month post-index was slightly more prevalent in the no-drain use group (3.30%, CI: 3.09%, 3.51%) compared to drain use group (3.01%, CI: 2.51%, 3.57%). When stratified by cemented or cementless fixation within the no drain group, cementless procedures had lower prevalence of general infection in both groups (No drain use: cemented= 2.76% CI: 2.48%, 3.06%, cementless=1.13%, CI: 0.37%, 2.61%).

Conclusions: Drain use during total knee replacement was not associated with postprocedural infections in unadjusted analyses, but was associated with lower surgeon procedural volume. Adjusted analyses are needed to control for potential confounding associations.