The Australian Experience: Evaluating Approach-Based Outcomes

Sarah O’Reilly-Harbidge

Introduction

The Australian population is relatively small compared with many other countries, being roughly equivalent to that of Texas or a third that of the United Kingdom despite having a landmass not dissimilar to the United States. It has been a long-standing rite of passage for many Australian surgeons to complete a period of international postfellowship training. This has resulted in a continued enthusiasm for broadening the local orthopaedic armamentarium for our patients’ benefit. This chapter briefly covers the Australian data, the national coordination of that data collection, research, and trends regarding the DAA to the hip. The discussion is focused on the notable contribution the AOANJRR has made to orthopaedics, with a specific early data focus on the DAA. The AOANJRR provides a large cohort of patient data that allows surgeons to compare and analyze the differences in outcomes between the three commonly used hip approaches (ie, anterior, lateral, and posterior).

Australian Orthopaedic Association National Joint Replacement Registry

Arthroplasty as a subspecialty has demonstrated a rapid evolution over the last century and is heavily technology driven. The AOANJRR is a comprehensive, nationally recognized arthroplasty database that was initially launched in 1998.¹ It has allowed time contemporaneous and constant assessment of this rapidly progressing field for the betterment of our patients receiving joint replacements. The registry was designed to prospectively collect data for primary and revision joint arthroplasty at the time of surgery from all hospitals across the country. This is achieved by using specific standardized forms (Figure 56.1). AOANJRR data collection forms are completed in the operating room at the time of surgery and are returned to the AOANJRR monthly. All procedures reported are cross-validated with independently collected state health department data in a sequential, multilevel matching process. In addition to this, data are also matched biannually with the national Department of Health and Ageing’s National Death Index to obtain information on the date of death where relevant.¹ This process enables near 100% data collection.

FIGURE 56.1 AOANJRR hip form.

A, side 1 and B, side 2.

(Reprinted with permission from the 2019 AOANJRR Annual Report.)

Approach to Data Analysis Within the Registry

Data from the AOANJRR allow us to predict the time to first revision of a prosthesis using the Kaplan-Meier estimates of survivorship. To avoid imprecise estimates at the right tail of the curve where the risk is low, the registry statisticians provide estimates until the proportion of prostheses at risk of revision is >10% of the initial number of cases up to 100; the graph is then extended. However, analytic comparisons of prostheses’ survival using log-rank tests and proportional hazards models are based on all available data.¹ These models are provided to aid in assessing overall differences in survival between both prosthesis types and surgical approach.

Given that there are many surgical- and patient- related factors that result in complex competing risks, the registry is investigating the best method of analysis to cope with this. To assess and mitigate competing risk, it requires a specific survival analysis that is designed to correctly estimate marginal probability of an event. The Kaplan-Meier product-limit method, the traditional method of assessing the data, does not mitigate multiple potential causes of the event outcome. Because of this, they may produce inaccurate estimates when assessing the marginal probability of cause-specific outcomes.

When considering the complexity of addressing competing risk analysis, the concept of cumulative incidence is one method of estimating the probability of revision. Cumulative incidence revision diagnosis graphs address the competing risks of reasons for revision and are specifically designed to highlight the differences between groups in the pattern of revision over time. This aids in the ability to assess to what extent the proposed variables influence the data, as well as looking at the mechanisms by which failure and revision occur.

Consideration must be given to patient death as an outcome because it affects the survivorship and risk of
revision data sets. Conditional probability is used to adjust the data set to include the probability of revision in patients who are alive at that particular point in time. If one were not to adjust for this, the risk would be driven down and the survivorship impossibly long. This is particularly pertinent when addressing arthroplasty in the elderly population.

Hip Arthroplasty Outcomes in Australia

The AOANJRR has evolved since its inception to include the full scope of joint arthroplasty and commenced data collection on hip arthroplasty operative approaches in 2015. Reporting of these data commenced in 2019, with a recent release of the first outcomes. For ease and accuracy, the approach has been divided into three anatomic descriptors: anterior, lateral, and posterior. There may be significant variation in the finessing of each of these broad categories by individual surgeons, although the principles behind each remain consistent. The data set includes reporting on the early outcomes for 122,345 total hip replacements, including 32,086 anterior (26.2%), 24,468 lateral (20.0%), and 65,791 posterior (53.8%) procedures for osteoarthritis,¹ with “early” defined as 5 years from the initial surgery. It is reasonable to expect that, with the incidence of total hip replacement for osteoarthritis estimated to rise by 208% by 2032 and with the increased interest in using the anterior approach, teasing out specific advantages and limitations of the approach in a large, prospectively acquired data set will be of great use to the orthopaedic community.

The extensive heterogeneity of the hip arthroplasty components that are available to the orthopaedic surgeon, both historically and contemporarily, means that the AOANJRR has limited the procedures analyzed for primary total hip replacement. Specifically, it is limited to procedures using modern bearings¹ (cross-linked polyethylene with metal, ceramic, or ceramicized metal heads and ceramic-on-ceramic bearings using mixed ceramic) in an effort to maintain relevance to modern surgical practice. Drawing from the established evidence, these bearings are associated with lower rates of revision, particularly in the long term, compared with other historically used bearings. There are marginal differences in outcomes between the different types of modern bearings. With this in mind, there is no current evidence to support the use of other bearing types.

Since 2014, 115,707 primary total hip replacements for osteoarthritis were included in the registry and the anterior approach was used in 26.7% of cases. The rate of revision is no different when the surgical approaches are compared. However, there were significant differences in the types of revision and the reasons for revision between the three approaches over the 5 years. The anterior approach has demonstrated a lower rate of revision for infection and dislocation; however, it has a higher rate of revision for aseptic loosening (0.4% anterior vs 0.2% lateral vs 0.1% posterior) and early fracture (0.6% anterior vs 0.3% for both posterior and lateral at 1 year).

The AOANJRR has provided analysis of the data performed with hazard ratios (HRs) adjusted for age, sex, American Society of Anesthesiologists score, body mass index (BMI) category, femoral fixation technique, and head size. The anterior approach has a higher rate of major revisions (HR = 2.1) compared with the posterior (HR = 1.6) and lateral approaches (HR = 1.6) out to 3 years; however, there was no difference between the posterior and lateral approaches. It should be noted that there is a higher rate of uncemented femoral prosthesis use in the anterior approach group; this is likely due to the greater availability of uncemented systems at the time, as well as the increased technical demand of cementing with anterior instrumentation. The current numbers show that the anterior approach is being used more often in a younger patient population than the posterior and lateral approaches. These patients also demonstrate a higher proportion of patients with both lower BMI and American Society of Anesthesiologists scores (Figure 56.2). Compared with the United States, where there is a strong uptake of uncemented femoral components, Australian surgeons have a more heterogeneous practice. The registry has identified subgroups within the data set that correlate with the type of prosthesis used. Those having a lower BMI or age at the time of operation more frequently undergo joint replacement using an uncemented prosthesis.

FIGURE 56.2 Primary total conventional hip replacement by age and BMI category and surgical approach (primary diagnosis osteoarthritis).

(Reprinted with permission from the 2019 AOANJRR Annual Report.)

Considering all three approaches, the leading indications for revision in the first 4 years are fracture, infection, dislocation, and loosening (Figures 56.3, 56.4, 56.5 and 56.6). The anterior approach has demonstrated a higher rate of revision for loosening and fracture compared with both the posterior and lateral approaches. There is also a significantly higher rate of revision in the first 3 months postoperatively in the anterior approach group. These complications are often attributed to unrecognized intraoperative undersizing of the femoral prosthesis with resultant uncontrolled subsidence or iatrogenic femoral fracture, both of which are often thought to be associated with what is described as the learning curve phenomenon, which is addressed later in the chapter.

Only gold members can continue reading. Log In or Register to continue