How Does Mortality Risk Change Over Time After Hip and Knee Arthroplasty?

Background Mortality after THA and TKA is lower than expected for several years after surgery when compared with age- and sex-adjusted population data. With long-term followup (beyond approximately 10 years), some evidence has suggested that this trend reverses, such that postsurgical mortality is higher than expected as more time passes. However, the degree to which this may be the case has not been clearly established.

Questions/purposes In this large-registry study, we asked: What is the long-term mortality after THA and TKA compared with the expected mortality, adjusted for age, sex, and calendar year.

Methods Using data on 243,057 THAs and 363,355 TKAs performed for osteoarthritis from the Australian Orthopaedic Association National Joint Replacement Registry (AOANJRR) from 2003 to 2016, and life tables from the Australian Bureau of Statistics the Standardised Mortality Ratio (SMR), relative mortality and excess mortality (relative to the expected mortality for people of the same sex and age in the same country) was calculated separately for hips and knees. The AOANJRR contains near-complete (98%-100%) data from all hospitals in Australia performing arthroplasty but does not include followup data on people who have left the country. Followup was from the date of surgery to 13 years, mean 5.8 years.

Results We found a lower-than-expected mortality for THA and TKA in the early years after surgery. This association diminished over time and the mortality became higher than expected after 12 years for both THA and TKA. For THA, the excess mortality (per thousand people) increased from 11 fewer deaths (95% CI, 10–11 fewer) after 1 year to four more deaths (95% CI, 0–9 more) in the 13th year, and the SMR increased from 0.50 (95% CI, 0.48–0.52) after 1 year to 1.07 (95% CI, 0.99–1.14) in the 13th year. For TKA, the excess mortality (per thousand people) increased from 12 fewer deaths (95% CI, 12–13 fewer) after 1 year to five more deaths (95% CI 2–9 more) in the 13th year, and the SMR increased from 0.39 (95% CI, 0.37–0.40) after 1 year to 1.09 (95% CI, 1.03–1.15) in the 13th year.

Conclusions Mortality after hip and knee arthroplasty is lower than expected (based on population norms) in the first 8 years to 9 years but gradually increases over time, becoming higher than expected after 12 years. The lower-than-expected mortality in the early years after surgery is likely the result of patient selection with patients undergoing primary arthroplasty having better health at the time of surgery than that of the age- and sex-matched population. The increasing mortality over time cannot be regression to the mean, as late mortality is higher than expected, moving beyond the mean. It is important to understand if there are modifiable factors associated with this increased mortality. The reasons for the change are uncertain. Factors to consider in future research include determining the effect of different patient factors on late mortality. Some of these included higher obesity rates for joint replacement patients and the association or causal impact of osteoarthritis and/or its treatment to increase late mortality in a similar manner to other forms of arthritis. There is also a possibility that the arthroplasty device itself may affect late mortality.

Level of Evidence Level III, therapeutic study.