Letters: Hip arthroplasty: Authors’ reply to Jameson and Breusch

Jameson and Breusch fail to appreciate that the adjusted hazard ratio, indicating higher mortality with cemented total hip replacement (THR), relates to all time points, including the first 60 days, with no indication that this hazard ratio varies over time.1 2 3

We repeatedly emphasised the potential for residual confounding and selection bias.1 Interestingly, when comparing revision rates, the National Joint Registry uses similar survival methods but adjusts for fewer confounding variables, so its data are also “flawed” by Jameson’s criteria.4 Revision comparisons are also prone to residual confounding and selection bias.

The Australian registry reports a hazard ratio of 1.82 between mortality rates in resurfacings and replacements.5 We found smaller but significant differences after adjusting for more confounders.

Pulmonary embolisation is known to cause pulmonary hypertension and long term right heart failure.6 Does non-thrombotic embolisation cause a similar problem? One study found that THR is associated with cerebral microemboli, whereas patients who have Birmingham hip resurfacing (BHR) are spared.7 The extra death within six years for every 23 cemented THRs compared with BHRs reported in our paper warrants further research.

It was not one of our pre-specified objectives to investigate BHRs by component size. Our analysis included only BHRs with 2 mm increment head sizes. Jameson also included BHRs with 4 mm increments.

Australian data show that implant survivorship at 10 years is 2% better for resurfacings than for THRs in all men with osteoarthritis under 65 years,8 contradicting Jameson’s assertion that indications for resurfacing are limited. Furthermore Wylde and Blom postulated a possible lower threshold for revision in resurfacing patients compared with THRs.9

We are surprised that Jameson thinks comparisons of mortality rates confuse patients. Patients are interested in their risk of death. Until we have better measures of comorbidity and activity, it is impossible to unravel the conundrum linking implant survival, mortality, activity,10 and healthy patient effects. Our paper does not mislead but highlights these discrepancies.