Is preoperative frailty a risk factor for poor outcomes after total joint replacements for osteoarthritis?

Abstract

Purpose: Frailty is a novel potentially modifiable risk factor for poor outcomes after surgical procedures. The objective of this study is to determine whether pre-operative frailty is associated with clinical outcomes 1-year after total hip or total knee replacement (TKR and THR) for osteoarthritis. Methods: Osteoarthritis patients ≥65yo scheduled for elective primary TKR or THR were recruited from a musculoskeletal specialty hospital. Frailty was measured using validated criteria comprised of specific frailty elements, including Katz Index of Independence in Activities of Daily Living (Katz ADL). Hip/Knee Injury and Osteoarthritis Outcome Score (HOOS/KOOS) were administered pre-operatively and at 1-year. Regression models were created by considering all variables which were significant at the 0.05 level in univariate models, and then performing backward selection to retain variables with 0.05 significance. Age and gender were forced in to all models. Results: 797 subjects enrolled, 325 THR/474 TKR. Median age 72 years (range 65-94), 98% Caucasian, 63.2% female, 7.0% were frail. 524 cases have submitted 1-year data, (94.4% of eligible). Frail subjects had significantly lower preoperative scores in all HOOS domains and in 3/5 KOOS domains. In a multivariable model the number of frailty elements pre-operatively was a significant predictor of OARSI-OMERACT responder status 1-year after joint replacement (OR = 1.48 all TJR; OR = 1.64 THR only; OR = 1.43 TKR only; all p-<0.05). In contrast, overall pre-operative frailty did not predict OARSI-OMERACT responder status. In the multivariable model pre-operative Katz ADL predicted 1-year HOOS Function in Daily living (β -10.9 +/- 2.6SD, p<0.001) and HOOS Function Sports and Rec. (β -17.0 +/- 5.1SD, p<0.001). The number of pre-operative frailty elements also predicted HOOS Function Sports and Rec. (β -4.20 +/- 1.9SD, p=0.02). In the univariate analysis, number of frailty elements significantly predicted severe adverse events in both TJR (OR = 1.34) and TKR (OR = 1.33) but this was no longer significant in the multivariable model. Overall frailty was not a significant predictor of adverse events in either univariate or multivariate analyses. Conclusions: Frail patients had significantly worse preoperative HOOS and KOOS scores. An increase in the number of frailty elements significantly increased the odds of being an OMERACT-OARSI responder at 1-year (OR = 1.43-1.64). Worse Katz ADL scores, an element of frailty, predicted worse HOOS function at 1-year. Frailty was not associated with serious adverse events, but a much larger cohort would be required to have adequate power to reliably assess this outcome. Monitoring the development of frailty or ADL disability may help inform optimal timing of TJR in OA patients contemplating this elective procedure.