Part Three: Outcomes after joint replacement 2003 to 2015
Part Three of the 13th Annual Report provides outcome data in relation to hip, knee, shoulder, elbow and ankle replacements. It describes activity between 1 April 2003 and 31 December 2015.
There were 2,055,687 procedures recorded in this period and 10% of these were excluded because there were insufficient patient details to enable linkage. This relates predominantly to the early years of the registry and was less of a feature in recent years as data quality has improved.
The numbers of primary procedures available for analysis were 796,636 total hip replacements, 871,472 knee replacements, 3,174 ankle replacements, 17,199 shoulder replacements and 1,631 elbow replacements.
Hip replacement procedures
The potential follow-up for hip procedures was 12.75 years. A total of 60% of the primary procedures were carried out on women and the median age at primary across the entire group was 69 years. Osteoarthritis was the predominant diagnosis in 92% of cases.
The most common form of fixation continues to be uncemented, but the percentage of total hip replacements that were uncemented has fallen to 39% from a peak of 46% in 2010. The trend for an increase in hybrid fixation seen over the last three reports has continued and now represents 26% of cases. The percentage of cemented total hip replacements performed has remained fairly static over the last seven years at just over 30%. Hip resurfacing remains at less than 1%. The most common articulation used in cemented, uncemented and hybrid prostheses continues to be metal-on-polyethylene. The trend in uncemented implantation showing a rise in ceramic-on- polyethylene and a decrease in ceramic-on-ceramic has continued with equal numbers now being used. With hybrid fixation, the increase in the use of ceramic-on- polyethylene reported last year has continued.
The Kaplan-Meier cumulative revision risk estimates are now reported at twelve years, with the lowest rates seen in the cemented fixation population at 3.93% (95% Confidence Interval 3.74-4.13), compared to 8.37% (95% CI 8.03-8.73) in the uncemented group. However, the uncemented group contained the majority of metal-on-metal articulations and when uncemented fixation was used with metal-on-polyethylene bearing, the twelve-year revision estimate was 5.46% (95% CI 4.92-6.06). The lowest revision risk in all categories was consistently seen with the ceramic-on-polyethylene bearing, the revision probability being 3.08% with the cemented fixation, 4.19% with uncemented and 3.29% with the hybrid fixation, although the latter was approximate as fewer than 250 were at risk at this point.
This year’s analysis continues to show the increased risk of revision associated with younger patients. For example, in female patients less than 55 years of age undergoing cemented hip replacement, the ten-year revision risk estimate was 5.85%, compared with 2.02% in females over 75 years. Similar trends are seen across all groups and gender, with an inverse relationship between the probability of revision and the age of the patient.
Our analysis of the relationship between head size and revision rates in hard-on-soft bearings (metal-on- polyethylene and ceramic-on-polyethylene) appears to indicate an ideal head size of between 26 and 32mm. Head sizes of 36mm and above are associated with increasingly higher failure rates.
The common stem brand combinations are reported in terms of revision risk, with further sub-division into bearing type. Several brands had low revision risk at ten years and were essentially comparable. The most commonly used cemented Exeter V40 with a Contemporary Flanged cup with metal-on-polyethylene bearing produced a ten-year revision estimate of 2.23%; the most widely used uncemented prosthesis the Corail Pinnacle with a metal-on-polyethylene bearing had a ten-year revision risk of 3.16% and the most widely used hybrid the Exeter V40 and uncemented Trident cup with metal-on-polyethylene bearing produced a ten-year revision risk of 2.75%. The ASR resurfacing had a revision estimate of 27.05% at ten years, rising to 30.35% at twelve years. (Note the twelve year figure is an approximation as fewer than 250 cases remained at risk).
The cumulative mortality was examined up to twelve years following primary surgery and as expected increased with age. For example, this was low in men under 55 years of age at 6.15% (95% CI 5.64-6.71) but rose to 94.32% (95% CI 92.08-96.10) in men over 85 years. The comparative figures are 5.96% (95% CI 5.40-6.58) and 85.97% (95% CI 84.22-81.52) for women in the same age groups.
The six most common indications for revision after primary total hip replacement (listed in order of frequency) remains aseptic loosening, pain, adverse soft tissue reaction to particulate debris, dislocation, infection and peri-prosthetic fracture. The rate of revision for aseptic loosening, pain and adverse soft tissue reaction to particulate debris tended to increase over time, reaching a maximum beyond five years. The rate of revision for dislocation, infection and peri-prosthetic fracture are at their highest within the first year following surgery.
The percentage of primary hip replacements performed for fractured neck of femur has increased gradually over the last twelve years reaching 4.5% in 2015 (3,733 procedures). Comparing the cohort of 19,872 primary hip replacements performed to treat fractured neck of femur, with those performed for all other causes showed a slightly higher revision risk and a greatly increased mortality risk at each time point in the fracture group.
Revision total hip replacement has been studied for data collected between 1 April 2003 and 31 December 2015. A total of 88,822 revision procedures were reported of which 87.2% were single-stage revisions, 6.0% were stage one of two-stage procedures and 6.8% were stage two of two-stage procedures. From 2003 to 2012 the number of revisions recorded annually increased from 1,426 in the first recorded year to 10,497 in 2012. Over the last three years there has been a reduction in numbers recorded to 8,367 in 2015.
The 88,822 revision procedures included multiple revision procedures entered for the same individual person-joint. Out of these, 78,130 first recorded revision procedures were identified for a given patient-side; 20,926 of these were revisions of primary operations that could be identified in the registry whilst the remaining 57,204 related to unrecorded primaries (either pre-dating 2003, the primary had not been captured in the NJR or the procedures could not be linked). The ten-year risk of re-revision following these first revision procedures was 14.83% (95% CI 14.38-15.31), which is approximately three times higher than the risk of revision in the primary cohort. The top five most common indications for re- revision (in order of greatest frequency) were aseptic loosening, dislocation/subluxation, infection, pain and peri-prosthetic fracture.
Knee replacement procedures
Of the 871,472 primary knee replacements, osteoarthritis was the sole stated indication for surgery in 96% of cases. Of all primary knee replacements, 84.7% were all cemented total knee replacements, the majority of which were unconstrained fixed bearing knees, 4.4% were uncemented and 1.0% were hybrid. The utilisation of unicondylar knee replacements remains similar to previous years at 8.7% of all procedures while patellofemoral replacement made up 1.3% of all procedures. A total of 57% of primary knee replacement surgeries were performed on women. The median age for a patient undergoing primary cemented total knee replacement surgery was 70 years and was 64 years for unicondylar replacement.
When considering the temporal change in implant selection between 2003 and 2015, the use of all cemented total knee replacement has risen from 81.5% of all recorded surgeries in 2003 to 87.4% in 2015.
There has been a decline in uncemented total knee replacements from 6.7% to 2.3% over the same time period. Unicondylar replacements remain between 8% and 9% of all primaries each year over the twelve-year period and patellofemoral replacements have continued to form just over 1% of all surgeries year on year.
The Kaplan-Meier cumulative revision risk estimates at twelve years were 3.82% (95% CI 3.71-3.94) for cemented total knee replacement, 4.74% (95% CI 4.34- 5.17) for uncemented total knee replacement and 4.17% (95% CI 3.47-5.00) for hybrid total knee replacement.
As reported in previous years the corresponding twelve-year revision estimate for unicondylar replacements were higher than total knee replacements at 14.99% (95% CI 14.16-15.87) and for patellofemoral replacement the revision risk was 23.83% (95% CI 21.19-26.73).
Revision estimates have been broken down according to level of constraint, for example the twelve-year estimate for cemented total knee replacement with an unconstrained, fixed bearing was 3.51% (95% CI 3.37-3.66) and the posterior-stabilised fixed bearing was 4.23% (95% CI 4.01-4.47). Further detailed breakdown in relation to fixation, bearing, constraint, gender and age show marked differences in outcomes. For example, when a cemented, unconstrained, fixed bearing total knee replacement was used in men over 75 years of age, the risk of revision at twelve years after the primary was just 2.14% (95% CI 1.79-2.56). In comparison, in men aged under 55, the revision risk estimate was 9.21% (95% CI 8.09-10.47).
The detailed breakdown of brands with a sub-division of fixation, bearing and constraint within brand continues to show that the ten-year revision estimates are low (less than 4% for many brands). For example looking at the most commonly used brands, at ten years the revision estimates were; 2.65% (95% CI 2.55-2.75) for the PFC Sigma, 3.62% (95% CI 3.43-3.82) for the Nexgen knee, 3.32% (2.83-3.90) for the Triathalon knee, 3.56% (95% CI 3.34-3.79) for the AGC and 2.78% (95% CI 2.51-3.07) for the Genesis 2.
Within the unicondylar brand group, the cumulative risk of revision at ten years varied from 6.31% (95% CI 5.16-7.70) seen with the Zimmer unicompartmental, to 12.02% (95% CI 11.51-12.54) with the Oxford prosthesis (the most commonly used) and 17.11% (95% CI 15.14- 19.32) with the Preservation.
The cumulative mortality at twelve years after the primary knee replacement for women under 55 years of age was 5.46% (95% CI 4.62-6.44) but rose to 85.79% (95% CI 83.27-88.10) in women over 85. The corresponding figures for male patients were 7.92 (95% CI 6.62-9.47) and 91.24 (95% CI 88.68-93.42).
Outcomes of revision knee replacement surgery are also reported. There were a total of 54,153 revision operations recorded in the NJR. In 2015, 79% of revisions were single-stage; 10.5% were stage one of two-stage and 10.5% were stage two of two-stage. Looking at the outcomes following the first revision recorded in NJR for a given patient-side, the twelve- year cumulative percentage probability of re-revision was 15.99% (95% CI 14.96-17.09). The re-revision risks were higher when the primary was recorded in the NJR at 16.76% (95% CI 15.66-17.92), compared to 14.19% (95% CI 13.07-15.39) when the primary was not recorded in the NJR.
Ankle replacement procedures
A total of 3,174 primary ankle replacements have been recorded in the NJR up to 31 December 2015. Ankle replacements were entered routinely from 2010 although 13 primary operations performed in 2008-2009 were entered. The 3,174 procedures were carried out by a total of 214 consultants in 228 hospitals. A total of 44% of consultants entered ten or more procedures over the five year period, which means that two-thirds of consultants are carrying out very small numbers per year. The maximum number carried out by any one unit was 234.
The median age at primary surgery remains at 68 years and 59% of procedures were carried out in men. A total of 94% of the procedures were uncemented.
The Mobility was the most commonly used brand of replacement until 2013, but it was withdrawn from the market in 2014. In 2015 the most commonly used prosthesis was the Zenith ankle (25.6%) followed by the Box ankle (22.3%) and the Infinity ankle (15.5%).
A total of 105 implantations have been revised and the five-year cumulative revision risk was 6.83% (95% CI 5.47-8.52).
Shoulder replacement procedures
A total of 17,199 primary shoulder replacements were recorded on the NJR from 1 April 2012 until 31 December 2015. These were carried out by a total of 636 surgeons in 369 units. The median number reported for each surgeon was 13 (IQR 2-41). The median age at primary surgery remains at 73 years and 71.4% of procedures were carried out in women.
Over the last year there has been a continued decrease in the use of resurfacing arthroplasty and an increased use of the reverse polarity total shoulder replacement, which, in 2015, represented over 45% of cases.
There were 364 shoulder revisions overall and the cumulative revision estimate at three years was 3.44% (95% CI 3.07-3.86). The relatively small numbers and short follow-up continues to prevent a detailed breakdown of causes of revision or differences between brands.
A detailed analysis of pre- and post-operative Patient Reported Outcome Measures has been undertaken on a sample of patients who had a primary shoulder replacement after April 2012. Of the total number of responses, 3,331 elective patients had completed both pre- and post-operative questionnaires. The median pre- operative Oxford Shoulder Score (OSS) was 16, rising to 36 at six months, with a median change score of 18. Overall 90.8% of the elective patients had improvement in their OSS, with 8.3% worse and 0.9% staying the same after surgery.