Mortality After Total Knee Arthroplasty: A Systematic Review of Incidence, Temporal Trends, and Risk Factors

Research output: Contribution to journalArticle (Academic Journal)peer-review

59 Citations (Scopus)
372 Downloads (Pure)


Background: To determine a contemporary estimate of the risk of mortality following total knee arthroplasty (TKA), including the identification of temporal trends, common causes, and modifiable and non-modifiable risk factors.

Materials and methods: Systematic review, meta-analysis and meta-regression. Searches of MEDLINE, AMED, CAB Abstracts, Embase, for studies in any language published between 2006 and 2016 reporting 30 or 90 day mortality following TKA, supplemented by contact with authors.

Results: Thirty-seven studies with mortality data from 15 different countries following 1.75 million TKAs contributed to this review. The pooled Poisson-normal random-effects meta-analysis estimate of 30 and 90 day mortality are 0.20% (95% confidence interval 0.17 to 0.24) and 0.39% (0.32 to 0.49). Both estimates have fallen over the study period (p<0.001). Meta-regression using median year of surgery as a moderator shows that 30 and 90 day mortality following TKA fell to 0.10% (0.07 to 0.14) and 0.19% (0.15 to 0.23) in 2015. The leading cause of death is cardiovascular disease.

Conclusions: There is a worldwide ongoing secular decline in mortality following TKA. Improved patient selection, perioperative care and a healthy population effect may account for this observation. Efforts to further reduce mortality should be targeted at reducing cardiovascular events following TKA.
Original languageEnglish
Pages (from-to)1064-1070
Number of pages7
JournalJournal of Bone and Joint Surgery
Issue number12
Early online date20 Jun 2018
Publication statusPublished - 20 Jun 2018

Structured keywords

  • Centre for Surgical Research


Dive into the research topics of 'Mortality After Total Knee Arthroplasty: A Systematic Review of Incidence, Temporal Trends, and Risk Factors'. Together they form a unique fingerprint.

Cite this