Abstract
Background
Long-term implant survivorship in THA and TKA involves a combination of factors related to the patient, the implants used, and the decision-making and technical performance of the surgeon. It is unclear which of these factors is the most important in reducing the proportion of revision surgery.
Questions/purposes
We used data from a large national registry to ask: In patients receiving primary THA and TKA for a diagnosis of osteoarthritis, do (1) the reasons for revision and (2) patient factors, the implants used, and the surgeon or surgical factors differ between surgeons performing THA and TKA who have a lower revision rate compared with all other surgeons?
Methods
Data were analyzed for all THA and TKA procedures performed for a diagnosis of osteoarthritis from the Australian Orthopaedic Association National Joint Replacement Registry (AOANJRR) from September 1, 1999, when collection began, to December 31, 2018. The AOANJRR obtains data on more than 98% of joint arthroplasties performed in Australia. The 5-year cumulative percent revision (CPR) was identified for all THAs and TKAs performed for a diagnosis of osteoarthritis with 95% confidence intervals (overall CPR); the 5-year CPR with 95% CIs for each surgeon was calculated for THA and TKA separately. For surgeons to be included in the analysis, they had to have performed at least 50 procedures and have a 5-year CPR. The 5-year CPR with 95% CIs for each THA and TKA surgeon was compared with the overall CPR. Two groups were defined: low revision rate surgeons (the upper confidence level for a given surgeon at 5 years is less than 3.84% for THA and 4.32% for TKA), and all other surgeons (any surgeon whose CPR was higher than those thresholds). The thresholds were determined by setting a cutoff at 20% above the upper confidence level for that class. The approach we used to define a low revision rate surgeon was similar to that used by the AOANJRR for determining the better-performing prostheses and is recommended by the International Prosthesis Benchmarking Working Group. By defining the groups in this way, a significant difference between these two groups is created. Determining a reason for this difference is the purpose of presenting the proportions of different factors within each group. The study group for THA included 116 low revision rate surgeons, who performed 88,392 procedures (1619 revised, 10-year CPR 2.7% [95% CI 2.6% to 2.9%]) and 433 other surgeons, who performed 170,094 procedures (6911 revised, 10-year CPR 5.9% [95% CI 5.7% to 6.0%]). The study group for TKA consisted of 144 low revision rate surgeons, who performed 159,961 procedures (2722 revised, 10-year CPR 2.6% [95% CI 2.5% to 2.8%]) and 534 other surgeons, who performed 287,232 procedures (12,617 revised, 10-year CPR 6.4% [95% CI 6.3% to 6.6%]). These groups were defined a priori by their rate of revision, and the purpose of this study was to explore potential reasons for this observed difference.
Results
For THA, the difference in overall revision rate between low revision rate surgeons and other surgeons was driven mainly by fewer revisions for dislocation, followed by component loosening and fracture in patients treated by low revision rate surgeons. For TKA, the difference in overall revision rate between low revision rate surgeons and other surgeons was driven mainly by fewer revisions for aseptic loosening, followed by instability and patellofemoral complications in patients treated by low revision rate surgeons. Patient-related factors were generally similar between low revision rate surgeons and other surgeons for both THA and TKA. Regarding THA, there were differences in implant factors, with low revision rate surgeons using fewer types of implants that have been identified as having a higher-than-anticipated rate of revision within the AOANJRR. Low revision rate surgeons used a higher proportion of hybrid fixation, although cementless fixation remained the most common choice. For surgeon factors, low revision rate surgeons were more likely to perform more than 100 THA procedures per year, while other surgeons were more likely to perform fewer than 50 THA procedures per year. In general, the groups of surgeons (low revision rate surgeons and other surgeons) differed less in terms of years of surgical experience than they did in terms of the number of cases they performed each year, although low revision rate surgeons, on average, had more years of experience and performed more cases per year. Regarding TKA, there were more differences in implant factors than with THA, with low revision rate surgeons more frequently performing patellar resurfacing, using an AOANJRR-identified best-performing prosthesis combination (with the lowest rates of revision), using fewer implants that have been identified as having a higher-than-anticipated rate of revision within the AOANJRR, using highly crosslinked polyethylene, and using a higher proportion of cemented fixation compared with other surgeons. For surgeon factors, low revision rate surgeons were more likely to perform more than 100 TKA procedures per year, whereas all other surgeons were more likely to perform fewer than 50 procedures per year. Again, generally, the groups of surgeons (low revision rate surgeons and other surgeons) differed less in terms of years of surgical experience than they did in terms of the number of cases they performed annually, although low revision rate surgeons, on average, had more years of experience and performed more cases per year.
Conclusion
THAs and TKAs performed by surgeons with the lowest revision rates in Australia show reductions in all of the leading causes of revision for both THA and TKA, in particular, causes of revision related to the technical performance of these procedures. Patient factors were similar between low revision rate surgeons and all other surgeons for both THA and TKA. Low revision rate THA surgeons were more likely to use cement fixation selectively. Low revision rate TKA surgeons were more likely to use patella resurfacing, crosslinked polyethylene, and cemented fixation. Low revision rate THA and TKA surgeons were more likely to use an AOANJRR-identified best-performing prosthesis combination and to use fewer implants identified by the AOANJRR as having a higher-than-anticipated revision rate. To reduce the rate of revision THA and TKA, surgeons should consider addressing modifiable factors related to implant selection. Future research should identify surgeon factors beyond annual case volume that are important to improving implant survivorship.
Level of Evidence
Level III, therapeutic study.
Introduction
National joint replacement registries can identify hip and knee implants (and the characteristics of implants, such as fixation or bearing type) that have a lower revision rate [9, 14, 18, 19]. Registries have emphasized that prosthesis selection is one of the most important factors affecting the revision rate [8]. Although implant selection is important, the decision-making process regarding implant choice and the surgical techniques used may influence THA and TKA revision rates more than the actual implants [11, 26]. Studies have found a relationship between a lower risk of revision after THA and TKA with various patient factors [24], such as age, sex, BMI, American Society of Anesthesiologists (ASA) class [11, 18], and surgeon factors (principally cases per year) [25]. Despite this, the revision rate has remained fairly constant [20, 27].
But there is limited research analyzing the variation in individual surgeons’ results and the patient, implant, and surgeon factors behind this difference, despite increasing reporting of surgeon variation in registry reports [2]. Survivorship of THA and TKA depends on a combination of patient, implant, and surgeon factors, is it is not clear which of these factors is most important. Identifying modifiable factors could lead to widespread improvements in survivorship [16, 23, 25]. Surgical performance cannot be assessed on a procedure-by-procedure basis using registry data, but surgical performance can be assessed by analyzing the results of surgeons with low revision rates after THA and TKA and comparing them with the results of all other surgeons. In doing so, the patient selection, prosthesis choice, and surgeon factors identified as contributing to increased survivorship can be used to modify surgical performance and improve patient outcomes. This has potential benefits to patients, surgeons, and healthcare systems.
We therefore used data from a large national registry to ask: In patients receiving primary THA and TKA for a diagnosis of osteoarthritis, do (1) the reasons for revision and (2) patient factors, the implants used, and the surgeon or surgical factors differ between surgeons performing THA and TKA who have a lower revision rate compared with all other surgeons?
Patients and Methods
Study Design and Setting
This is a comparative study between low revision rate surgeons and all other surgeons, performed in the context of a large national arthroplasty registry. The Australian Orthopaedic Association National Joint Replacement Registry (AOANJRR) began collecting data on September 1, 1999. Registry data are validated against patient-level data provided by each of the State and Territory Health Departments in Australia with the use of a sequential, multilevel matching process. A matching program is run monthly to search for all primary and revision arthroplasties recorded in the AOANJRR that involved the same side and joint for the same patient, thus enabling each revision to be linked to the primary procedure. Data were also matched by the Australian Institute of Health and Welfare National Death Index to obtain information on the date of death. After cross-checking data, the registry is able to obtain data on more than 98% of joint arthroplasties performed in Australia [3]. The AOANJRR records the reasons for revision and the type of revision THA and TKA.
Procedures undertaken by known surgeons performing primary conventional THA and TKA for osteoarthritis from September 1, 1999 to December 31, 2018 were included. Procedures with metal-on-metal THA were excluded because of their known higher revision rates. The 5-year cumulative percent revision (CPR) was identified for all THAs and TKAs performed for a diagnosis of osteoarthritis with 95% confidence intervals (overall CPR), and the 5-year CPR with 95% CIs for each surgeon was calculated for THA and TKA separately. For surgeons to be included in the analysis, they had to have performed at least 50 procedures and have a 5-year CPR. The 5-year CPR with 95% confidence intervals for each THA and TKA surgeon was compared with the overall CPR. Two groups were defined: low revision rate surgeons (the upper confidence level for a given surgeon at 5 years is less than 3.84% for THA and 4.32% for TKA) and all other surgeons (any surgeon whose CPR was higher than those thresholds). The thresholds were determined by setting a cutoff at 20% above the upper confidence level for that class. Our approach to defining a low revision rate surgeon was similar to that used by the AOANJRR for determining the better-performing prostheses, which is based on the upper confidence level. This approach also is recommended by the International Prosthesis Benchmarking Working Group [13]. By defining the surgeon groups this way, a significant difference between these two groups is created. Determining a reason for this difference is the purpose of presenting the proportions of different factors within each of these groups. Any further statistical analysis will be an artifact of the difference between groups created by this definition. Therefore, a multivariate analysis is not possible, and only descriptive statistics are provided.
For each of these defined groups, we compared patient factors, including age, gender, BMI, and ASA class. For THA, we compared implant factors, including implant choice, number of combinations used, fixation, bearing surface, head size, and articulation. For TKA, we compared implant choice, number of combinations used, fixation, bearing, patella resurfacing, use of navigation, and robotic use. For each group we also compared surgeon factors, including surgeon experience (years of follow-up), volume (cases per year), location of practice, and surgical approach to THA.
Implant choice was assessed according to the surgeon’s use of prosthesis combinations reported as either the “best performing” (combinations reported in the AOANJRR annual report with the lowest rates of revision) or as having a higher-than-anticipated rate of revision (combinations reported in the AOANJRR annual report with the highest rates of revision) [4]. The best-performing THA prosthesis combinations (Supplementary Table 1; http://links.lww.com/CORR/A648) were defined as those with a 5-year CPR ≤ 1.8% and at least 500 procedures, and the best-performing TKA prosthesis combinations (Supplementary Table 2; http://links.lww.com/CORR/A649) were defined as those with a 5-year CPR ≤ 2.3% for TKAs and at least 500 procedures. The AOANJRR identifies higher-than-anticipated rate of revision THA and TKA prosthesis combinations using a standardized three-stage approach. Higher-than-anticipated rate of revision combinations are initially identified if the revision rate (per 100 component years) exceeds twice that of all other prostheses in the same class and the Poisson probability of observing that number of revisions, given the proportion of the class, is statistically significant (p < 0.05) [3, 4].
For THA, the overall CPR at 5 years was 3.1% (95% CI 3.0% to 3.2%) (Fig. 1). The study group consisted of 116 low revision rate surgeons performing 88,392 procedures (1619 revised) and 433 other surgeons performing 170,094 procedures (6911 revised) (Supplementary Fig. 1; http://links.lww.com/CORR/A650). Low revision rates surgeons had a 5-year CPR of 1.4% (95% CI 1.4% to 1.5%) and 10-year CPR of 2.7% (95% CI 2.6% to 2.9%). All other surgeons had a 5-year CPR of 3.7% (95% CI 3.6% to 3.8%) and 10-year CPR of 5.9% (95% CI 5.7% to 6.0%) (Supplementary Table 3; http://links.lww.com/CORR/A651). There were 1518 recorded THA prosthesis combinations.
Fig. 1.
This graph shows the cumulative percent revision for conventional THA by surgeon group; LRR = low revision rate. A color image accompanies the online version of this article.
For TKA, the overall CPR at 5 years was 3.5% (95% CI 3.5% to 3.6%) (Fig. 2). The study group consisted of 144 low revision rate surgeons performing 159,961 procedures (2722 revised) and 534 other surgeons performing 287,232 procedures (12,617 revised) (Supplementary Fig. 2; http://links.lww.com/CORR/A652). Low revision rate surgeons had a 5-year CPR of 1.6% (95% CI 1.6% to 1.7%) and 10-year CPR of 2.6% (95% CI 2.5% to 2.8%). All other surgeons had a 5-year CPR of 4.4% (95% CI 4.3% to 4.5%) and 10-year CPR of 6.4% (95% CI 6.3% to 6.6%) (Supplementary Table 4; http://links.lww.com/CORR/A653). There were 264 recorded TKA prosthesis combinations.
Fig. 2.
This graph shows the cumulative percent revision for conventional TKA by surgeon group; LRR = low revision rate. A color image accompanies the online version of this article.
The study groups were defined a priori by their revision rate, and the purpose of this study was to explore potential reasons for this observed difference.
Primary and Secondary Study Endpoints of Interest
Our primary study goal was to identify the differences in the reasons for revision between low revision rate surgeons and all other surgeons. To achieve this, we identified all revision THAs and TKAs performed by our defined two groups and compared the difference in the CPR for each factor. Our secondary study goal was to compare patient, implant, and surgeon factors between low revision rate surgeons and all other surgeons. To achieve this, a comparison of the descriptive statistics was made as a result of our method of defining the two study groups.
Ethical Approval
The AOANJRR is approved by the Commonwealth of Australia as a federal quality assurance activity under section 124X of the Health Insurance Act, 1973. All AOANJRR studies are conducted in accordance with ethical principles of research (the Helsinki Declaration II).
Results
Differences in Reasons for Revision
When the reasons for THA revision were analyzed, low revision rate surgeons had reductions in the four most common complications (Fig. 3): loosening (Fig. 4A), fracture (Fig. 4B), dislocation (Fig. 4C), and infection (Fig. 4D).
Fig. 3.
These graphs show the cumulative incidence revision diagnosis for conventional THA by surgeon group; LRR = low revision rate. A color image accompanies the online version of this article.
Fig. 4.
A-D These graphs show the cumulative percent revision of primary conventional THA for a primary diagnosis of osteoarthritis by surgeon group revised for (A) loosening, (B) fracture, (C) dislocation, and (D) infection; LRR = low revision rate. A color image accompanies the online version of this article.
When the reasons for TKA revision were analyzed, low revision rate surgeons had reductions in the five most common complications (Fig. 5): component loosening (Fig. 6A), infection (Fig. 6B), patella complications (Fig. 6C), instability (Fig. 6D), and pain (Fig. 6E).
Fig. 5.
These graphs show the cumulative incidence revision diagnosis for TKA by surgeon group; LRR = low revision rate. A color image accompanies the online version of this article.
Fig. 6.
A-E These graphs show the cumulative percent revision of primary TKA for a primary diagnosis of osteoarthritis by surgeon group revised for (A) loosening, (B) infection, (C) patella reasons, (D) instability, and (E) pain; LRR = low revision rate. A color image accompanies the online version of this article.
Differences in Patient, Implant, and Surgeon Factors
Patient factors were generally similar between low revision rate surgeons and all other surgeons for THA, with similar proportions of men (46%), age groups, ASA class, and BMI (Table 1). Regarding implant factors, low revision rate surgeons used half as many prosthetic combinations collectively, but they averaged slightly more implant combinations per surgeon. Low revision rate surgeons used a higher proportion of hybrid fixation (38%) than other surgeons (28%), but cementless fixation remained the most common fixation choice in both groups (58% low revision rate surgeons, 66% other surgeons). Metal-on–highly crosslinked polyethylene was the most-used bearing surface in both groups, followed by ceramic-on-ceramic and ceramic-on–crosslinked polyethylene. Low revision rate surgeons were less likely to use a higher-than-anticipated rate of revision prosthesis than were other surgeons. Both groups had a similar proportion of procedures with each head size. Groups were similar regarding articulation and selection of a best-performing prosthetic combination.
Table 1.
Differences in patient, implant, and surgeon factors between LRR surgeons performing THA and all other surgeons
| Variable | LRR surgeons % (n = 88,392) |
All other surgeons % (n = 170,094) |
| Men | 46 (40,554 of 88,392) | 46 (78,279 of 170,094) |
| Age in years | ||
| < 55 | 11 (9935 of 88,392) | 11 (17,959 of 170,094) |
| 55-64 | 25 (21,852) | 24 (40,653) |
| 65-74 | 36 (31,418) | 36 (60,982) |
| 75 or older | 28 (25,187) | 30 (50,500) |
| ASAa | ||
| ASA class 1 | 12 (5395 of 43,862) | 10 (8640 of 85,734) |
| ASA class 2 | 57 (25,130) | 54 (46,694) |
| ASA class 3 | 29 (12,825) | 34 (29,074) |
| ASA class 4 | 1.2 (512) | 1.5 (1316) |
| ASA class 5 | 0 (0) | 0 (10) |
| BMIb | ||
| Underweight | 0.7 (196 of 29,512) | 0.8 (428 of 55,874) |
| Normal | 23 (6806) | 20 (11,439) |
| Preobese | 39 (11,376) | 37 (20,653) |
| Obese class 1 | 24 (7042) | 26 (14,401) |
| Obese class 2 | 9 (2753) | 11 (6048) |
| Obese class 3 | 5 (1339) | 5 (2905) |
| Fixation | ||
| Cemented | 4 (3659 of 88,392) | 5 (9165 of 170,094) |
| Cementless | 58 (51,316) | 66 (112,717) |
| Hybrid | 38 (33,417) | 28 (48,212) |
| Head size in mm | ||
| < 32 | 18 (15,873 of 88,392) | 19 (33,012 of 170,094) |
| 32 | 40 (35,538) | 38 (63,982) |
| > 32 | 42 (36,981) | 43 (73,100) |
| Bearing surface | ||
| Ceramic-on-ceramic | 32 (28,577 of 88,392) | 23 (39,932 of 170,094) |
| Ceramic-on-non-XLPE | 0.3 (306) | 2 (3866) |
| Ceramic-on-XLPE | 23 (19,959) | 20 (34,676) |
| Metal-on-non-XLPE | 5 (4131) | 6 (10,540) |
| Metal-on-XLPE | 34 (29,924) | 41 (69,620) |
| Ceramicized metal-on-XLPE | 6 (5402) | 7 (11,072) |
| Other | 0.1 (93) | 0.2 (388) |
| Surgical approachb | ||
| Anterior | 29 (8856 of 30,114) | 25 (14,543 of 58,149) |
| Lateral | 17 (5263) | 22 (12,549) |
| Posterior | 53 (15,995) | 53 (31,057) |
| Articulation | ||
| Constrained | 0.1 (123 of 88,392) | 0.1 (244 of 170,094) |
| Dual-mobility | 1.0 (846) | 1.9 (3247) |
| All other prostheses | 99 (87,423) | 98 (166,603) |
| Type of hospital | ||
| Public | 20 (18,049 of 88,392) | 25 (42,845 of 170,094) |
| Private | 80 (70,343) | 75 (127,249) |
| Identified HTARR prosthesesc | ||
| Identified combination or prostheses (procedures) | 4 (3442 of 88,392) | 13 (21,582 of 170,094) |
| Not identified (procedures) | 96 (84,950) | 87 (148,512) |
| Identified combination or prostheses (comb) | 23 (158 of 688) | 26 (337 of 1299) |
| Number of combinations | ||
| Best combinations (procedures) | 6 (5000) | 2 (2969) |
| Other combinations (procedures) | 94 (83,392) | 98 (167,125) |
| Surgeon volume, average per year | ||
| < 10 procedures | 0.6 (554 of 88,392) | 3 (5505 of 170,094) |
| 11-25 procedures | 6 (5582) | 19 (33,081) |
| 26-50 procedures | 27 (23,449) | 38 (64,987) |
| 51-100 procedures | 38 (33,725) | 35 (60,083) |
| 100+ procedures | 28 (25,082) | 4 (6438) |
| Surgeon experience | ||
| < 10 years of procedures in the AOANJRR | 3 (2304 of 88,392) | 9 (15,696 of 170,094) |
| 10+years of procedures in the AOANJRR | 97 (86,088) | 91 (154,398) |
Data presented as % (n) of procedures; “best” was defined as procedure combinations with the lowest 5-year cumulative percent revision that were used in more than 500 procedures; 116 LRR surgeons performed 88,392 hip procedures; 433 other surgeons performed 170,094 hip procedures.
ASA has only been collected by the AOANJRR since 2012.
BMI and surgical approach has only been collected by the AOANJRR since 2015.
Percentage of the number of prosthesis combinations are displayed in this section; LRR = low revision rate; ASA = American Society of Anesthesiologists; XLPE = crosslinked polyethylene; HTARR = higher than anticipated rate of revision.
Regarding surgeon factors, low revision rate surgeons were more likely to perform more than 100 THA procedures per year, whereas all other surgeons were more likely to perform between 25 and 50 THA procedures per year. Low revision rate surgeons had a higher proportion of those with at least 10 years’ experience, but the groups were similar as to whether the operation was performed in a private or public hospital or which surgical approach was used.
Patient factors were generally similar between low revision rate surgeons and all other surgeons for TKA, with similar proportions of men, age groups, ASA class, and BMI (Table 2). Regarding implant factors, low revision rate surgeons used fewer prosthetic combinations collectively but averaged slightly more implant combinations per surgeon. Low revision rate surgeons were more likely to resurface the patella, use a better-performing prosthesis combination, not use a higher-than-anticipated rate of revision prosthesis, use highly crosslinked polyethylene, and were less likely to use cementless fixation compared with the other surgeons. The amount of implant constraint was similar between the two groups, as was the use of computer navigation, image-derived instrumentation, and robot use. Regarding surgeon factors, low revision rate surgeons were more likely to perform more than 100 TKA procedures per year, and all other surgeons were more likely to perform between 25 and 50 TKA procedures per year. Low revision rate surgeons were slightly more experienced than all other surgeons, and the groups were similar as to whether the operation was performed in a private or public hospital.
Table 2.
Differences in patient, implant, and surgeon factors between LRR surgeons performing TKA and all other surgeons
| Variable | LRR surgeons % (n = 159,961) |
All other surgeons % (n = 287,232) |
| Men | 44 (70,636 of 159,961) | 44 (126,165 of 287,232) |
| Age in years | ||
| < 55 | 6 (8836 of 159,961) | 7 (20,620 of 287,232) |
| 55-64 | 26 (41,432) | 28 (79,272) |
| 65-74 | 41 (65,005) | 39 (112,003) |
| 75 or older | 28 (44,688) | 26 (75,337) |
| ASAa | ||
| ASA grade 1 | 6 (5403 of 83,611) | 7 (9515 of 144,230) |
| ASA grade 2 | 59 (49,063) | 55 (79,479) |
| ASA grade 3 | 34 (28,437) | 37 (53,469) |
| ASA grade 4 | 0.8 (701) | 1 (1760) |
| ASA grade 5 | 0.0 (7) | 0.0 (7) |
| BMIb | ||
| Underweight | 0.2 (105 of 55,031) | 0.2 (172 of 95,552) |
| Normal | 11 (6165) | 10 (9785) |
| Preobese | 33 (18,323) | 31 (29,694) |
| Obese class 1 | 30 (16,629) | 31 (29,512) |
| Obese class 2 | 16 (8631) | 17 (16,182) |
| Obese class 3 | 9 (5178) | 11 (10,207) |
| Fixation | ||
| Cemented | 58 (92,799 of 159,961) | 56 (160,987 of 287,232) |
| Cementless | 16 (24,861) | 22 (62,225) |
| Hybrid | 26 (42,301) | 22 (64,020) |
| Patella resurfacing | ||
| Patella used | 69 (110,404 of 159,961) | 53 (152,808 of 287,232) |
| No patella used | 31 (49,557) | 47 (134,424) |
| Stability | ||
| Fully stabilized | 0.2 (396 of 159,961) | 0.4 (1182 of 287,232) |
| Hinged | 0.1 (135) | 0.2 (461) |
| Medial pivot design | 3 (5430) | 3 (9915) |
| Minimally stabilized | 72 (114,489) | 67 (193,436) |
| Posterior stabilized | 25 (39,487) | 29 (82,170) |
| Unknown | 0.0 (24) | 0.0 (68) |
| Type of hospital | ||
| Public | 22 (35,900 of 159,961) | 25 (70,929 of 287,232) |
| Private | 78 (124,061) | 75 (216,303) |
| Polyethylene usec | ||
| Non-XLPE | 54 (86,663 of 159,910) | 65 (188,056 of 287,160) |
| XLPE | 46 (73,247) | 35 (99,104) |
| Navigation use | ||
| Computer-navigated | 23 (37,305 of 159,961) | 24 (68,736 of 287,232) |
| Non-navigated | 77 (122,656) | 76 (218,496) |
| Image-derived instrumentation (IDI) | ||
| IDI used | 7 (11,701 of 159,961) | 6 (17,197 of 287,232) |
| No IDI | 93 (148,260) | 94 (270,035) |
| Robotic use | ||
| Robotically assisted | 0.5 (805 of 159,961) | 0.5 (1555 of 287,232) |
| Nonrobotic | 100 (159,156) | 99 (285,677) |
| Identified HTARR prosthesesd | ||
| Identified combination or prostheses (procedures) | 2 (3425 of 159,961) | 10 (29,418 of 287,232) |
| Not identified (procedures) | 99 (156,536) | 90 (257,814) |
| Identified combination or prostheses (combinations) | 31 (54 of 177) | 23 (58 of 251) |
| Number of combinations | ||
| Best combinations (procedures) | 29 (46,017 of 159,961) | 13 (38,286 of 287,232) |
| Other combinations (procedures) | 71 (113,944) | 87 (248,946) |
| Surgeon volume, average per year | ||
| < 10 procedures | 0.2 (288 of 159,961) | 1 (3442 of 287,232) |
| 11-25 procedures | 1 (2168) | 9 (26,366) |
| 26-50 procedures | 10 (16,421) | 29 (81,865) |
| 51-100 procedures | 39 (62,654) | 45 (129,017) |
| 100+ procedures | 49 (78,430) | 16 (46,542) |
| Surgeon experience | ||
| < 10 years of procedures in the AOANJRR | 4 (6642 of 159,961) | 10 (29,828 of 287,232) |
| 10+ years of procedures in the AOANJRR | 96 (153,319) | 90 (257,404) |
Data are presented as % (n) knee procedures; “best” was defined as combinations with the lowest 5-year cumulative percent revision that were used in more than 500 procedures; 144 LRR surgeons performed 159,961 hip procedures; 534 other surgeons performed 287,232 hip procedures.
ASA has only been collected by the AOANJRR since 2012.
BMI has only been collected by the AOANJRR since 2015.
Patients with unknown bearing surfaces were excluded.
Percentage of the number of prosthesis combinations are displayed in this section; LRR = low revision rate; ASA = American Society of Anesthesiologists; XLPE = crosslinked polyethylene; IDI = image-derived instrumentation; HTARR = higher-than-anticipated rate of revision.
Discussion
Survivorship of THA and TKA depends on patient, implant, and surgeon variables. There are a group of surgeons with lower revision rates than all other surgeons, and it is unknown which factors pertain to the low revision rates and produce the variation in results. Identifying these factors may benefit all surgeons as well as the patients undergoing arthroplasty procedures. Using data from the AOANJRR, we identified surgeons with the lowest THA and TKA revision rates and compared their reasons for revision with all other surgeons; we also compared the patient, implant, and surgeon factors between these two surgeon groups. We found that low revision rate surgeons had reductions in all the main causes of THA and TKA revision, in particular reasons of revision related to the technical performance of the procedure. For THA, the difference in overall revision rate between low revision rate surgeons and other surgeons was driven mainly by fewer revisions for dislocation, followed by component loosening and fracture in patients treated by low revision rate surgeons. For TKA, the difference in overall revision rate between low revision rate surgeons and other surgeons was driven mainly by fewer revisions for aseptic loosening, followed by instability and patellofemoral complications in patients treated by low revision rate surgeons. Patient factors were generally similar between groups. Differences associated with implant choice were more evident for TKA, but also were apparent for THA. Surgeon volume affected the revision rate for both THA and TKA. Surgeons should consider modifying the implant factors identified. It remains to be seen whether additional training, theoretical knowledge, improved judgement, and/or a higher case volume alone will reduce the revision rate. Future research should enhance the type of data collected for patient and surgeon factors and should aim to identify other surgeon-related factors (aside from case volume) that can improve arthroplasty survivorship.
Limitations
The largest limitation of the study is the approach to creating the study groups. Using CIs to highlight low revision rate surgeons (applying a similar method used by registries to identify implants that have a higher-than-expected revision rate), we have created a statistically different group of surgeons from the remaining surgeons. This means that when identifying the patient, implant, and surgeon factors that are different between the low revision rate surgeons and all other surgeons, we are unable to perform a statistical analysis and can only provide descriptive statistics. Some of the identified differences may not be independent of each other, and we cannot conclusively say which factors are the most important for improving survivorship. Univariate and multivariate analyses may identify which factors are the most significant in improving overall results, but they require a different study design. As far as we are aware, there are no similar studies that have attempted to identify all possible factors from a registry to differentiate the results of low revision rate surgeons from all other surgeons. This study should be used for generating further hypotheses and alternate study designs. This will allow better determination of whether the revision risk is reduced through improved patient selection, using better-performing implants, or through better surgical decision-making and techniques. Another limitation is that factors related to surgical technique and performance cannot be easily quantified or measured in a registry (and some are perhaps not measurable at all). These are likely to have an impact on the overall survivorship. There are several factors that may be important in producing variance in survivorship results. For example, we were unable to assess patient factors beyond gender, age, ASA class, and BMI. We could not study specific patient comorbidities that are known to affect surgical outcome, such as diabetes, diabetic control, smoking, malnutrition, renal impairment, and other organ failure [1, 22]. Although a crude measure, the ASA score reflects a patient’s overall degree of comorbidity. Scoring systems such as the Charlson comorbidity index may provide more informative data. Surgeon factors including the level of subspecialty training, arthroplasty fellowship completion, location of training, and success with fellowship examinations are not recorded. Case complexity in terms of severity of arthritis, preoperative alignment, bone loss, ROM, and prior surgeries are also not recorded, nor are alignment goals for TKA, case judgement, or rationale for surgical decision-making.
Another limitation is patient selection and the degree of procedural complexity and/or the influence of registry results on the decision-making process. More complex procedures may typically be performed by surgeons who are better trained, more experienced, and who perform surgery in higher volume, but not always. Case complexity may lead to an increased revision rate, but alternatively, in the hands of some surgeons it may not. Without being able to identify preoperative case complexity, this cannot be determined. Some surgeons may be more restrictive with patient and procedure selection, which itself may reduce revision rates and complications. Without access to individual case details and radiographs, a registry study is unable to analyze this. An opt-in process to record pre- and postoperative radiographs may allow for this to be assessed. The publication of registry results may impact patient and case selection, with some surgeons choosing to avoid complex cases to produce lower revision rates. This could be more important in the future or in countries where individual surgeon results are publicly released, where there is strong influence from online reviews, or where there are financial penalties with complications, such as in bundled payment models. In Australia, this is not the case, and most arthroplasty procedures are performed in private hospital settings. We believe it is unlikely to have influenced our results, but such limitations in practice may prove to be important variables in reducing the overall revision burden.
Finally, there are limitations to extrapolating this data outside the Australian population. Alternate factors or trends may be identified in other national registries. This study only examined revision rates and did not assess patient-reported outcomes, patient satisfaction, or a combination of measures. It is unclear whether there is a correlation between surgeons with low revision rates and these outcomes.
Differences in Reasons for Revision
The most obvious reasons for the reduction in revision for THA were dislocation, aseptic loosening, and fracture. Dislocation is a leading cause of THA revision [7]. Attempts to reduce the frequency of dislocation have largely been achieved with changes in femoral head size, articulations, and surgical approach. Furthermore, enabling technologies are being promoted. Although larger head sizes, dual mobility constructs, and constrained acetabular liners have reduced dislocations [6, 10, 12], we found no difference in femoral head size or articulations between low revision rate surgeons and other surgeons. Dual mobility constructs were used by low revision rate surgeons in only 1% of procedures. We also found no difference in the surgical approach between low revision rate surgeons and other surgeons. The anterior approach has a reduced rate of dislocation compared with the posterior approach [12], but some surgeons might be able to achieve comparable results regardless of surgical approach used. There are factors beyond femoral head size, articulations, and surgical approach that are not included in this study that contribute to a higher risk of dislocation, and these are likely to be related to the technical performance [10]. Identifying these patient and surgeon factors may help reduce the revision burden. Likewise, the reductions in component loosening and fracture could be related to the increased selective use of femoral component cement fixation through the use of hybrid fixation and not using an identified prosthesis with a known higher revision rate and/or improvements in surgical technique. Technical errors during femoral preparation and implant insertion, rather than the implant used, are likely important factors in the early revision differences between the groups. The revision burden has only gradually decreased for THA, even though multiple international registries have produced annual reports and focused studies have aimed to reduce revision [20]. Enhanced data collection through data linkages and registry-nested clinical trials may identify other important patient, surgeon, and technique factors that reduce the revision rate.
The most common reasons for the reduction in TKA revision were aseptic loosening, instability, and patellofemoral complications. The reduction in aseptic loosening could be partly attributed to not using prostheses with a known higher revision rate, a higher proportion of cemented fixation, and the use of better-performing prostheses. It is unclear how surgeon factors relate, such as quality of the cement mantle achieved for fixation and the overall implant alignment produced. The reduction in revision due to instability could be related to bearing selection because the level of constraint was similar between groups. There was an obvious difference in the proportion of patella resurfacing performed and revisions for patellofemoral complications. Patellofemoral resurfacing is seemingly an obvious way to reduce revision for this complication, but it is unlikely to have an impact on the other modes and causes of revision. Given the international variation in patella resurfacing [5, 9], other countries may identify other differences between low revision rate surgeons and all other surgeons using a similar research design.
Differences in Patient, Implant, and Surgeon Factors
We did not identify differences in patient factors that may reduce revision rates, and this may be related to the limited dataset used in this study. Enhanced data collection through data linkages and registry-nested clinical trials may identify other important patient, surgeon, and technique factors that reduce the revision rate. We suggest not using implants that have been identified as having a higher-than-anticipated revision rate and to increase selective cement usage through hybrid fixation. It has been shown in a previous registry study that appropriate implant choice reduces revision surgery [8]. The most obvious differences identified between low revision rate surgeons and other surgeons were surgeon factors, particularly volume. Annual case volume has been associated with a reduced frequency of complications [15], and surgical technique may also be responsible for improved surgical outcomes [25]. Hospital volume has been considered regarding its relationship to arthroplasty complications [21]. However, we found no difference between procedures performed in public versus private hospitals. Future research should better investigate the differences in surgeon factors given the different complication profiles and the similarities in many implant factors. Future studies should also assess the benefit of arthroplasty centers, staffed with arthroplasty fellowship-trained surgeons with a special interest in arthroplasty who perform high-volume care, which currently is not the standard of care in Australia.
Similar to THA, we did not identify patient factors that reduce TKA revision rates. For implant factors, low revision rate surgeons appear to more closely follow registry-identified practice associated with lower revision rates, including increased patella resurfacing, use of highly cross-linked polyethylene, using a better-performing prosthesis, and not using an identified prosthesis with higher-than-anticipated revision rates. Surgeon volume was one of the most obvious differences between low revision rate surgeons and other surgeons. Surgeon volume has been shown to reduce adverse events [16] and has been associated with reduced infection, procedure time, length of stay, transfusion rate, and improved patient-reported outcomes [17]. Higher surgeon volume results in more accurate implant positioning, which may affect survivorship [16], but other factors related to surgeon volume leading to improved results must be further assessed. Implant factors were more related than with THA. Future research should investigate whether surgeons who modify their implant choices observe reduced revision rates; future studies might also further investigate surgeon factors related to the technical aspects of these procedures, given the different complication profiles identified. The role of enabling technologies in improving the results for surgeons with a higher revision rate needs to be assessed, although we have identified no difference in usage of navigation, image-derived instrumentation, or robotics between low revision rate surgeons and all other surgeons.
Conclusion
Surgeons with a low THA revision rate mostly achieve this by reducing revisions caused by dislocation, aseptic loosening, and fracture. Surgeons with a low revision rate for TKA mostly achieve this by reducing revisions for aseptic loosening, instability, and patellofemoral complications. Patient-related differences were not apparent between groups. Low revision rate surgeons used fewer implants that have been identified with a higher-than-anticipated revision rate. For THA, low revision rate surgeons used selective hybrid fixation more often. For TKA, low revision rate surgeons perform patella resurfacing, used better performing prostheses, highly crosslinked polyethylene, and cemented fixation more often. The most obvious difference between low revision rate THA and TKA surgeons and all other surgeons was the number of cases performed per year, with higher volume surgeons having fewer revisions. To reduce the revision rate, arthroplasty surgeons should consider addressing the modifiable implant factors that have been identified. A practical approach may be for complex procedures and procedures in younger patients (who, presumably, have a higher revision risk) to be performed by higher volume surgeons with lower revision rates. Auditing modifiable prosthesis variables for surgeons with a higher revision rate may help, too. Registry studies have predominantly focused on implant factors, but enhanced data collection may allow the identification of other important patient, surgeon, and technique factors that reduce the revision rate.
Supplementary Material
Acknowledgments
We thank the AOANJRR, as the data source used in this study, for the generous assistance provided in the development of the study methodology and data analysis.
Footnotes
Each author certifies that there are no funding or commercial associations (consultancies, stock ownership, equity interest, patent/licensing arrangements, etc.) that might pose a conflict of interest in connection with the submitted article related to the author or any immediate family members.
All ICMJE Conflict of Interest Forms for authors and Clinical Orthopaedics and Related Research® editors and board members are on file with the publication and can be viewed on request.
Ethical approval for this study was not sought.
This work was performed at the Australian Orthopaedic Association National Joint Replacement Registry, based at the South Australian Health and Medical Research Institute, Adelaide, Australia.
Contributor Information
Sophia Rainbird, Email: srainbird@aoanjrr.org.au.
Michelle Lorimer, Email: Michelle.Lorimer@sahmri.com.
Stephen E. Graves, Email: segraves@aoanjrr.org.au.
Roger Bingham, Email: rogerbingham@mac.com.
References
- 1.Alamanda V, Springer B. The prevention of infection: 12 modifiable risk factors. Bone Joint J. 2019;101:3-9. [DOI] [PubMed] [Google Scholar]
- 2.Australian Orthopaedic Association National Joint Replacement Registry. Hip, knee & shoulder arthroplasty: annual report 2017. Available at https://aoanjrr.sahmri.com/documents/10180/397736/Hip%2C%20Knee%20%26%20Shoulder%20Arthroplasty. Accessed November 14, 2020.
- 3.Australian Orthopaedic Association National Joint Replacement Registry. Hip, knee & shoulder arthroplasty: annual report 2019. Available at https://aoanjrr.sahmri.com/documents/10180/668596/Hip%2C+Knee+%26+Shoulder+Arthroplasty/c287d2a3-22df-a3bb-37a2-91e6c00bfcf0. Accessed November 14, 2020.
- 4.de Steiger RN, Miller LN, Davidson DC, Ryan P, Graves SE. Joint registry approach for identification of outlier prostheses. Acta Orthop. 2013;84:348-352. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Fraser JF, Spangehl MJ. International rates of patellar resurfacing in primary total knee arthroplasty, 2004-2014. J Arthroplasty. 2017;32:83-86. [DOI] [PubMed] [Google Scholar]
- 6.Garbuz DS, Masri BA, Duncan CP, et al. The Frank Stinchfield Award: Dislocation in revision THA: do large heads (36 and 40 mm) result in reduced dislocation rates in a randomized clinical trial? Clin Orthop Relat Res. 2012;470:351-356. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Gwam CU, Mistry JB, Mohamed NS, et al. Current epidemiology of revision total hip arthroplasty in the United States: National Inpatient Sample 2009 to 2013. J Arthroplasty. 2017;32:2088-2092. [DOI] [PubMed] [Google Scholar]
- 8.Harris I, Cuthbert A, Lorimer M, et al. Outcomes of hip and knee replacement surgery in private and public hospitals in Australia. ANZ J Surg. 2019;89:1417-1423. [DOI] [PubMed] [Google Scholar]
- 9.Heckmann N, Ihn H, Stefl M, et al. Early results from the American Joint Replacement Registry: a comparison with other national registries. J Arthroplasty. 2019;34:S125-S134. e121. [DOI] [PubMed] [Google Scholar]
- 10.Hoskins W, Bingham R, Hatton A, de Steiger RN. Standard, large-head, dual-mobility, or constrained-liner revision total hip arthroplasty for a diagnosis of dislocation: an analysis of 1,275 revision total hip replacements. J Bone Joint Surg Am. 2020;102:2060-2067. [DOI] [PubMed] [Google Scholar]
- 11.Hoskins WT, Bingham RJ, Lorimer M, de Steiger RN. The effect of size for a hydroxyapatite-coated cementless implant on component revision in total hip arthroplasty: an analysis of 41,265 stems. J Arthroplasty. 2020;35:1074-1078. [DOI] [PubMed] [Google Scholar]
- 12.Hoskins W, Bingham R, Lorimer M, Hatton A, de Steiger RN. Early rate of revision of total hip arthroplasty related to surgical approach: an analysis of 122,345 primary total hip arthroplasties. J Bone Joint Surg Am. 2020;102:1874-1882. [DOI] [PubMed] [Google Scholar]
- 13.International Prosthesis Benchmarking Working Group. International prosthesis benchmarking working group guidance document: hip and knee arthroplasty devices. Available at: Report by International Prosthesis Benchmarking Working Group May 2018 Final Draft (003).pdf - Google Drive. Accessed June 15, 2021.
- 14.Jameson SS, Asaad A, Diament M, et al. Antibiotic-loaded bone cement is associated with a lower risk of revision following primary cemented total knee arthroplasty: an analysis of 731 214 cases using national joint registry data. Bone Joint J. 2019;101:1331-1347. [DOI] [PubMed] [Google Scholar]
- 15.Jolback P, Rolfson O, Cnudde P, et al. High annual surgeon volume reduces the risk of adverse events following primary total hip arthroplasty: a registry-based study of 12,100 cases in Western Sweden. Acta Orthop. 2019;90:153-158. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Kazarian GS, Lawrie CM, Barrack TN, et al. The impact of surgeon volume and training status on implant alignment in total knee arthroplasty. J Bone Joint Surg Am. 2019;101:1713-1723. [DOI] [PubMed] [Google Scholar]
- 17.Lau RL, Perruccio AV, Gandhi R, Mahomed NN. The role of surgeon volume on patient outcome in total knee arthroplasty: a systematic review of the literature. BMC Musculoskelet Disord. 2012;13:250. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Lenguerrand E, Whitehouse MR, Beswick AD, et al. Risk factors associated with revision for prosthetic joint infection following knee replacement: an observational cohort study from England and Wales. Lancet Infect Dis. 2019;19:589-600. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Lindberg-Larsen M, Jorgensen CC, Solgaard S, et al. Increased risk of intraoperative and early postoperative periprosthetic femoral fracture with uncemented stems. Acta Orthop. 2017;88:390-394. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.McGrory BJ, Etkin CD, Lewallen DG. Comparing contemporary revision burden among hip and knee joint replacement registries. Arthroplast Today. 2016;2:83-86. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Mufarrih SH, Ghani MOA, Martins RS, et al. Effect of hospital volume on outcomes of total hip arthroplasty: a systematic review and meta-analysis. J Orthop Surg Res. 2019;14:1-13. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Podmore B, Hutchings A, van der Meulen J, Aggarwal A, Konan S. Impact of comorbid conditions on outcomes of hip and knee replacement surgery: a systematic review and meta-analysis. BMJ Open. 2018;8:e021784. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Price AJ, Alvand A, Troelsen A, et al. Knee replacement. Lancet. 2018;392:1672-1682. [DOI] [PubMed] [Google Scholar]
- 24.Schwartz FH, Lange J. Factors that affect outcome following total joint arthroplasty: a review of the recent literature. Curr Rev Musculoskelet Med. 2017;10:346-355. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Skoogh O, Tsikandylakis G, Mohaddes M, et al. Contemporary posterior surgical approach in total hip replacement: still more reoperations due to dislocation compared with direct lateral approach? An observational study of the Swedish Hip Arthroplasty Register including 156,979 hips. Acta Orthop. 2019;90:411-416. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Vertullo C, Lewis P, Lorimer M, Graves S. The effect on long term survivorship of surgeon preference for posterior stabilized or minimally stabilized total knee replacement: an analysis of 63416 cases from the AOANJRR. Orthop J Sports Med. 2017;5:2325967117S2325900170. [DOI] [PubMed] [Google Scholar]
- 27.Wilson S, Marx RG, Pan T-J, Lyman S. Meaningful thresholds for the volume-outcome relationship in total knee arthroplasty. J Bone Joint Surg Am. 2016;98:1683-1690. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.