When should we do knee replacements?

Abstract

It is not known whether it is better to perform total knee replacements early or late in the course of arthritis. To examine the determinants of patient-related outcomes after knee replacement, we examined 812 patients in the Avon knee registry, in whom pre-operative and 5-year post-operative American Knee Society Scores were available. Ninety percent had a good outcome at 5 years. Women had significantly more severe disease at the time of surgery than men. Older people (>79 years) had the worst disease at the time of surgery, and made the least health gain, although their overall outcomes were still good. Those with the most severe disease at the time of operation made greater gains to those with less disease, but remained in worse health at 5 years, suggesting that earlier surgery may be preferable.

Introduction

Knee replacement surgery has had a dramatic effect on the treatment of severe knee arthritis [1], [2]. A variety of recent reviews have concluded that both unicompartmental and total knee replacements (TKRs) are valuable and cost-effective [3], [4], [5]. Operation rates are rising in the UK [6], but there is some evidence for a large unmet need for these operations [1], [7], and with the ageing of the population it is likely that demand will increase [8].

There are at least three problems with these procedures. First, knee replacement is not totally safe, having a significant incidence of peri- and post-operative mortality and morbidity. Second, these operations are expensive, and the relative cost-effectiveness of surgery, in comparison with conservative interventions, has never been tested. Thirdly, there are a proportion of people who do not obtain the great benefits in pain and function that follow surgery in the majority, or in whom the prosthesis fails after a relatively short time [9]. Given these problems, it is vital that we learn how to maximise the value of knee replacement surgery, by carrying it out at a time at which any given individual would be most likely to obtain the most benefit. There are surprisingly few studies of these matters. It has been suggested that earlier total joint replacements may decrease the length of stay and prevent loss of function [10], [11]. However, we remain largely ignorant as to which patients are most likely to do well or badly, or when to operate.

Timing of surgery is the key issue. Surgeons need to know the optimum window of opportunity in relation to the age of the patient, their disease duration and the severity of the joint disease. The available consensus based criteria for indications and prioritisation do not answer these questions [12], [13]. Analysis of large observational databases (registries) provides one opportunity to examine the outcomes of people having TKRs at different times. The larger the database, the more varied the patients at the time of surgery, and the longer the period of observation, the more likely it is that the data derived will be both generalisable and reliable.

The Bristol Knee Replacement Registry was established in 1975 by the Bristol Knee Group. It contains information on a total of 2032 patients coming to primary knee surgery between 1974 and 1998. Survivorship data of the main prostheses used has already been published [14], [15]. The purpose of the present study was to examine this database in order to relate demographic data at the time of surgery to the change in health status.