Response to two recent BMJ papers on mortality after hip replacement: comparative modelling study

Two BMJ observational studies report increased survival of patients after metal-on-metal resurfacing versus cemented or uncemented total hip replacement.1 2 We considered the potential usefulness of these new estimates for lifetime cost effectiveness models, which necessitate extrapolation beyond observed data.3 Cost effectiveness models have assumed similar mortality irrespective of implanted device.4 5 6

McMinn and colleagues examined men under 55 years who received Birmingham hip resurfacing or cemented or uncemented hip replacements (mean age ~48 years).1 Kendal and colleagues compared resurfacing with cemented and uncemented procedures in propensity matched groups followed for 10 years (mean age 57-59 years; 55-59% men).2 McMinn and colleagues’ estimated six year survival after resurfacing was 0.995; survival in those who underwent cemented and uncemented procedures were 0.018 and 0.013 worse, respectively.

The figure⇓ shows Office of National Statistics mortality data and well fitting Gompertz distribution for 48 year old men. Keeping the shape parameter we calculated the scale parameters that generated McMinn and colleagues’ six year estimates. Plotted over a lifetime, these generated the curves in the top part of the figure⇓, which predicted slightly better survival compared with the general population after uncemented total hip replacement, slightly worse after cemented a procedure, whereas after Birmingham hip resurfacing about one third of patients are predicted to become centenarians. The bottom part of the figure⇓ shows the curves generated when applying the same procedure with Kendal and colleagues’ 10 year data. Again, survival after resurfacing is remarkably good, with a fifth of patients predicted to be centenarians.