Background: The capacity for total knee arthroplasty to improve pain, quality of life, and functional outcomes is widely recognized. Postoperative mortality is rare but of paramount importance, and needs to be accurately quantified and conveyed to patients in order to support decision-making prior to surgery. The purpose of this study was to determine a contemporary estimate of the risk of mortality following total knee arthroplasty, including the identification of temporal trends, common causes, and modifiable and nonmodifiable risk factors.
Methods: We performed a systematic review with searches of MEDLINE, AMED, CAB Abstracts, and Embase. Studies in any language published from 2006 to 2016 reporting 30 or 90-day mortality following total knee arthroplasty were included, supplemented by contact with authors. Meta-analysis and meta-regression were performed for quantitative data.
Results: Thirty-seven studies with mortality data from 15 different countries following over 1.75 million total knee arthroplasties formed the basis of this review. The pooled Poisson-normal random-effects meta-analysis estimates of 30 and 90-day mortality were 0.20% (95% confidence interval [CI], 0.17% to 0.24%) and 0.39% (95% CI, 0.32% to 0.49%). Both estimates have fallen over the 10-year study period (p < 0.001). Meta-regression using the median year of surgery as a moderator showed that 30 and 90-day mortality following total knee arthroplasty fell to 0.10% (95% CI, 0.07% to 0.14%) and 0.19% (95% CI, 0.15% to 0.23%), respectively, in 2015. The leading cause of death was cardiovascular disease.
Conclusions: There is an ongoing worldwide temporal decline in mortality following total knee arthroplasty. Improved patient selection and perioperative care and a healthy-population effect may account for this observation. Efforts to further reduce mortality should be targeted primarily at reducing cardiovascular events following total knee arthroplasty.
Level of evidence: Therapeutic Level IV. See Instructions for Authors for a complete description of levels of evidence.