If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Department of Trauma and Orthopaedic Surgery, Manchester University NHS Foundation Trust, Manchester, UKDepartment of Orthopedic Surgery, Yitzhak Shamir Medical Center (Assaf Harofeh), affiliated to the Sackler Faculty of Medicine, Tel-Aviv University, Tel-Aviv, Israel, Zerifin, Israel
Global Center of Medical Engineering and Informatics, Osaka University, Osaka, JapanOrthopaedic Surgery, Osaka University Graduate School of Medicine, Osaka, Japan
Department of Orthopaedic Surgery MassGeneralBrigham (MGB), Brigham and Women’s Distinguished Chair in Orthopaedic Surgery Chief: Division of Sports Medicine Director: Cartilage Repair Center Brigham and Women’s Hospital Harvard Medical School, Chestnut Hill, Massachusetts, USA
Trauma and Orthopaedics, Liverpool University Hospitals NHS Foundation Trust, Liverpool, UKDivision of Cell Matrix Biology and Regenerative Medicine, Faculty of Biology, Medicine and Health, School of Biological Sciences, The University of Manchester, Manchester, UK
Sports injuries, trauma and the globally ageing and obese population require increasing levels of knee surgery. Shared decision making has replaced the paternalistic approach to patient management. Evidence-based medicine underpins surgical treatment strategies, from consenting an individual patient to national healthcare system design. The evolution of successful knee-related registries starting from specific arthroplasty registries has given rise to ligament reconstruction, osteotomy and cartilage surgery registries developing as platforms for surgical outcome data collection. Stakeholders include surgeons and their patients, researchers, healthcare systems, as well as the funding insurers and governments. Lately, implant manufacturers have also been mandated to perform postmarket surveillance with some hoping to base that on registry data.
Aiming to assess the current status of knee-related registries, we performed a comprehensive literature and web search, which yielded 23 arthroplasty, 8 ligament, 4 osteotomy and 3 articular cartilage registries. Registries were evaluated for their scope, measured variables, impact and limitations.
Registries have many advantages as they aim to increase awareness of outcomes; identify trends in practice over time, early failing implants, outlier surgeon or institution performance; and assist postmarketing surveillance. International collaborations have highlighted variations in practice. The limitations of registries are discussed in detail. Inconsistencies are found in collected data and measured variables. Potential measurement and selection biases are outlined. Without mandated data collection and with apparent issues such as unverified patient reporting of complications, registries are not designed to replace adverse event recording in place of a proper safety and efficacy study, as demanded by regulators. Registry ‘big data’ can provide evidence of associations of problems. However, registries cannot provide evidence of causation.
Hence, without careful consideration of the data and its limitations, registry data are at risk of incorrectly drawn conclusions and the potential of misuse of the results. That must be guarded against.
Looking at the future, registry operators benefit from a collective experience of running registries as they mature, allowing for improvements across specialties. Large-scale registries are not only of merit, improving with stakeholder acceptance, but also are critical in furthering our understanding of our patients’ outcomes. In doing so, they are a critical element for our future scientific discourse.
Patient registries have proven themselves as an important tool for observational studies. The use of registries enables physicians, researchers and pharmaceutical and implant manufacturers to assess patient, drug and medical devices-related information.
Registries have several advantages over other study designs, namely, their large size, real-world data appeal and potential for long-term follow-up. In addition, they allow for critical analyses such as clinical/cost-effectiveness assessment, evaluation of failing implants and thus, postmarket surveillance of implants and techniques, which has increasingly been mandated in our field.
The first orthopaedic registries were hospital-based or regional-based databases. The Leeds Bone Tumour Registry was the earliest published database found in our search of the literature, established in 1958.
Their goal was to evaluate the revision outcomes of knee arthroplasties. Following the realisation of the opportunity and further improvements in the process, many countries and organisations have created new registries, not only for joint arthroplasty but also for other techniques (anterior cruciate ligament reconstruction (ACLR)) or other anatomical sites (spine, shoulder, foot and ankle). These registries help to contribute to better assessment of surgical procedures, patient selection, implant design and economic data.
While registries aim to increase awareness for the importance of outcomes, they are particularly powerful tools to identify outliers and trends in practice over time. They support evidence-based practice and enable some elements of monitoring postmarketing surveillance. Thus, registries are powerful tools to help patients, clinicians, health systems and commercial manufacturers in their diagnostic and treatment-related decision-making processes. Moreover, registries can also serve as the historical control of new products and interventions as alternative to randomised controlled trials (RCTs).
However, registries have recognised weaknesses too. While registry data are effective at identifying trends, registry studies are observational studies with inherent limitations. First, registries may take several years to reach high enough enrolment numbers to give representative results. This might lead to early unreliable reports with both false-positive and false-negative associations. Second, registry-based studies evaluate associations and not causation. Hence, they ‘point the finger to a problem’ but do not explain the exact reason why a problem happens. The identification of trends should trigger relevant hypotheses that then should be tested.
Despite their limitations, registries allow for the collection and interpretation of valuable epidemiological, demographic, economic and outcome measures related data characterising different aspects of surgical or non-surgical treatment.
The importance of registries has been widely accepted in the last couple of decades as it has been recognised that it serves patients, healthcare professionals, commercial manufacturers and government/national authorities. However, despite the potential advantages, establishing a successful registry is not an easy task, and many issues should be taken into consideration when planning the structure of the database. According to Kolling et al,
a successful registry should be measured by several key factors which include the way it is organised (funding, maintenance, methods of data collection, etc); documentation (which is measured by a validated scoring system and completeness of more than 90% of the target population data); and data handling and the scientific output of a register (both publications and periodic reports). Although their report was aimed to assess national arthroplasty registries, their key points are also pertinent to other registries.
The objective of this paper was to review the current status of knee-related registries, discuss their scope, measured variables, impact and limitations. We focused on knee arthroplasty, ligament, osteotomy and cartilage injury registries.
Methods
The membership list of the International Society of Arthroplasty Registries
) were reviewed in order to identify the most influential knee arthroplasty registries. Ligament, osteotomy and cartilage registries were identified using web searching using the key words “knee”, “registry or registries”, “ligament or ACL”; “osteotomy or osteotomies”; “cartilage or chondral”. These were validated through a search on PubMed, Embase, Web of Science and the Cochrane Library databases (figure 1). Inactive registries were excluded from this review. Each registry was evaluated using its annual report, official websites and publications in peer-reviewed journals (table 1). Several variables were extracted to assess the structure of the registry, its maturity as a functional registry and its potential impact. These variables are summarised in Table 2, Table 3, Table 4, Table 5, Table 6, Table 7, Table 8, Table 9. PubMed, Embase, Web of Science and the Cochrane Library databases were also reviewed to identify publications which demonstrate the use and impact of these registries.
Figure 1Literature search process. ISAR, International Society of Arthroplasty Registries; NORE, Network of Orthopaedic Registries of Europe.
Table 2Arthroplasty registry general information: name, type, year established, number of primary and revision total knee arthroplasties, implant type, fixation type, indication for revision and link to website/publication/annual report
Regional Register of Orthopaedic Prosthetic Implantology Report of R.I.P.O. Overall data hip, knee and shoulder arthroplasty in Emilia-Romagna region (Italy) 2000-2017;.
Table 3Ligament registry general information: name, type, year established, number of primary ACLRs, graft type, fixation type, follow-up time, number of revisions, incidence of concomitant injuries and link to website/publication/annual report
Table 4Osteotomy registry general information: name, type, year established, number of primary osteotomies, location of osteotomy, number of reoperations, indication for reoperation and link to website/publication/annual report
Table 5Cartilage injury registry general information: name, type, year established, aetiology, location, size, number of primary cartilage restoration surgeries, restoration modality, follow-up time, number of revisions, incidence of concomitant injuries and link to website/publication/annual report
Degenerative isolated cartilage defects of the patellofemoral joint are associated with more severe symptoms compared to trauma-related defects: results of the German cartilage registry (KnorpelRegister DGOU).
Table 6Variables reported by arthroplasty registry: demographics, revision rate/survival, surgical complications, medical complications, mortality, use of prophylactic antibiotics, thrombosis prophylaxis, PROMs, radiological assessment, healthcare use and healthcare costs
Regional Register of Orthopaedic Prosthetic Implantology Report of R.I.P.O. Overall data hip, knee and shoulder arthroplasty in Emilia-Romagna region (Italy) 2000-2017;.
EQ-5D, EuroQol Five-Dimensional; KOOS, Knee Injury and Osteoarthritis Outcome score; KOOS JR, Knee Injury and Osteoarthritis Outcome Score for Joint Replacement; KOOS-PS, KOOS Physical Function Shortform; N/A, not applicable; NRS, Numeric Rating Scale; OKS, Oxford Knee Score; PROM, patient-reported outcome measure; PROMIS-10, Patient-Reported Outcomes Measurement Information System; SF-12, 12-Item Short Form Health Survey ; SF-36, 36-Item Short Form Health Survey; VAS, visual analogue score; VR-12, The Veterans RAND 12 Item Health Survey; WOMAC, Western Ontario and McMaster Universities Osteoarthritis Index.
Table 8Variables reported by osteotomy registry: demographics, revision rate/survival, surgical complications, medical complications, mortality, use of prophylactic antibiotics, thrombosis prophylaxis, PROMs, radiological assessment, healthcare use and healthcare costs
Table 9Variables reported by cartilage injury registry: demographics, revision rate/survival, surgical complications, medical complications, mortality, use of prophylactic antibiotics, thrombosis prophylaxis, PROMs, radiological assessment, healthcare use and healthcare costs
Degenerative isolated cartilage defects of the patellofemoral joint are associated with more severe symptoms compared to trauma-related defects: results of the German cartilage registry (KnorpelRegister DGOU).
Data from each included registry were extracted and organised in tables according to the registry focus—arthroplasty, ligament, osteotomy and cartilage. Extracted data included name of registry, catchment area (national, regional and global), date established, number of pathways for specified indications for primary and revision surgeries according to registry focus (eg, implant type/fixation for arthroplasty registries, graft type for ACLR, and lesion size and location for cartilage injuries). Furthermore, the most recently published reports were reviewed for availability of measured variables. Where an official report was unavailable or when the report was written in a language other than English, peer-reviewed publications were analysed. These variables include demographics (gender, age and Body Mass Index (BMI)), surgical outcomes and complications, patient-reported outcome measures (PROMs), use of prophylactic antibiotics and thromboprophylaxis. The complete lists of these findings are presented in Table 6, Table 7, Table 8, Table 9.
Knee arthroplasty registries
The first national knee arthroplasty registry was established in Sweden in 1975. The other Scandinavian countries developed theirs during the 1980s and 1990s. There are many national knee arthroplasty registries, 10 of which met our inclusion criteria and are evaluated in this paper. Most registries allow open access to their annual reports. However, some reports are only available for local physicians (including the Scandinavian registries), and evaluation of their data was made possible as they are included in peer-reviewed publications.
Some hospital-based and regional registries were used as the foundations for later established national registries. The Geneva Arthroplasty Registry is a good example for this process.
Besides its participation in the Swiss Arthroplasty Registry, it continues to operate as a stand-alone registry with many variables which were not included in most of the national registries. These include additional patient information such as comorbidities and patient activity.
Except for the Scottish and Romanian national registries, all registries recorded implant design/model and method of fixation (cemented, cementless or hybrid). The indication for revision surgery was well documented in most registries. A comparison between the recording of other variables is detailed in Table 2, Table 6. One important finding is the inconsistency in PROM assessment. The majority of registries did not record PROMs and used revision as the endpoint. The registries which did are summarised in table 6. This is also true for recording of prophylactic antibiotic administration and thrombosis prophylaxis use. However, lacking PROMs in their last available reports, some registries reported that they have started or will start collection of PROMs in the near future (ie, the Norwegian, Canadian, New Zealand and Australian registries).
Patient-related factors such as age and sex were recorded in all reviewed registries. BMI was recorded in most registries. However, other variables such as socioeconomic status, comorbidities, medical complications, and perioperative adjuvant treatment such as antibiotics and thromboprophylaxis use were documented inconsistently. Regarding the latter two issues, we found the Portuguese registry to be unique among national registries because of its recording of many important variables which are more characteristic of smaller regional/hospital registries. It collects comorbidities, level of primary surgeon and assistant, surgical approach, patella intervention, general and regional anaesthesia details, financial coverage and more.
Ligament registries
Ligament registries are younger than arthroplasty registries. The first recorded ligament registry was the Multicenter Orthopaedic Outcomes Network in 2002. The Scandinavian registries followed in 2004–2005. Currently published postoperative follow-up times are between 1 and 3 years. Some registries declare the intention of continuing follow-up until 10 years postoperatively. As of yet, the Swedish Ligament Registry is the only registry to achieve this goal. The extracted data are summarised in Table 3, Table 7.
All the reviewed registries and studies recorded data on concomitant injuries. Ligament registries in general assess PROMs, an improvement from many of the arthroplasty registries, but there is no uniformity in the PROMs chosen. As with the arthroplasty registries, inconsistencies exist with regard to perioperatively and intraoperatively collected data. While several registries report postoperative radiological assessment, none of the registries require documentation of the tunnel position at both the femoral and tibial sides. Such documentation can improve the registry’s data quality, as malposition is known to be a critical determinant of postoperative laxity.
The relationship between the angle of the tibial tunnel in the coronal plane and loss of flexion and anterior laxity after anterior cruciate ligament reconstruction.
In general, data on healthcare use or healthcare costs are lacking.
A unique feature reporting a stated termination plan after 10 years of follow-up for its registry duration has been a design feature of the Multi-Centre ACL Revision Study
There has been a recent resurgence in interest in this procedure particularly for younger osteoarthritic (OA) patients’ treatment. Purely osteotomy-focused registries have been developed during the past decade, with some of them still in their early stages of development. This manifests in limited reports and relatively low numbers of registry-based publications in peer-reviewed journals. The Finnish National Hospital Discharge Register (FNHDR) is one of the oldest registries, starting data collection in 1956 and achieving nationwide coverage in 1967.
In 2012, Niinimäki et al described osteotomy surgery records in his registry, although a registry that was not osteotomy focused. Nevertheless, they identified a total of 6004 osteotomies (5734 patients) between the years 1987 and 2008.
In 2013, The SKAR used its well-established platform and prospectively started collecting data from patients who underwent a knee osteotomy. Their first report was published in 2014,
and their most recent report which summarised 2018, reported 163 new records, reaching a total record of over 934 (despite the 2016 annual report being unavailable for analysis). Their report includes comprehensive data of patient characteristics, including prior knee surgeries, indication for surgery and perioperative data, including approach, type of fixation, type of anaesthesia, operating time and prophylactic treatments (Table 4, Table 8). In 2014, the UK established a separate registry for osteotomies—the UK Knee Osteotomy Registry (UKKOR).
This is the largest osteotomy-focused registry in the world to date. In their annual report from 2018, they reported data from 49 surgeons who contributed 1652 patients (a total of 1776 procedures), of which 621 had already undergone surgery.
and was approved to add a knee osteotomy pathway by the local federal authorities. Their most recent annual report did not state the number of osteotomies and their data were not available for this review.
There are few cartilage registries with only three accessible operative registries. The extracted data are summarised in Table 5, Table 9. Only one registry is operated at a national level—the Danish National Patient Register (DNPR). However, despite the fact that over 22 254 cases of primary cartilage repair or reconstruction are recorded in the DNPR (1996–2011
), it is not a cartilage-oriented registry and consequently has some methodological limitations. One example is the potential underestimation of the incidence of cartilage injury. The DNPR is based on the documentation of surgical procedures in which cartilage injuries were addressed. This could lead to an overestimation of the proportion of high-grade injuries as many low-grade injuries are treated non-operatively and may be missed.
The other two registries were established more recently. The German Cartilage Registry (KnorpelRegister DGOU, German Society for Orthopaedics and Trauma Surgery)
Degenerative isolated cartilage defects of the patellofemoral joint are associated with more severe symptoms compared to trauma-related defects: results of the German cartilage registry (KnorpelRegister DGOU).
is a multicentre registry that was founded in 2013. This registry is based on a web-based remote data entry system from 23 cartilage repair centres in Germany, Austria and Switzerland, with up to 72 patients entered per month. As a multicentre registry, it has many advantages including records of lesion size, location, and aetiology. It holds records of both preoperative MRI for lesion assessment and postoperative MRI for regenerated cartilage assessment. It collects information regarding the surgical technique, the objective and subjective outcomes, including revision rates.
One drawback of this registry is the lack of information regarding concomitant injuries and perioperative complications.
The International Cartilage Regeneration and Joint Preservation Society (ICRS) global cartilage registry was launched in 2016. It is a multilingual, multinational registry, collecting data from surgeons of over 50 countries. It was designed to focus on cartilage injury, its natural history and management outcomes. As opposed to the other cartilage registries, it collects data about both low-grade and high-grade lesions which are treated either surgically or non-surgically. It records lesion location and modality of treatment. Other collected data include concomitant injuries, operative complications and use of prophylactic antibiotics. It does not report MRI data. This registry records EuroQol Five-Dimensional that will eventually lead to the ability to compare cost-effectiveness of differing techniques.
The UK’s National Institute for Health and Care Excellence recommends that clinicians collect and enter data from patients undergoing mosaicplasty onto this registry.
National Institute for Health and Care Excellence (NICE) Mosaicplasty for symptomatic articular cartilage defects of the knee. Interventional procedures guidance IPG607;.
The Swiss Medical Regulation Authority has mandated data collection onto the ICRS registry for all patients treated using Spherox, a product containing spheroids of human autologous matrix-associated chondrocytes for implantation.
Our search process is described in figure 1. Database search yielded 4620 papers on knee registries: 2669 arthroplasty registry papers, 426 ligament registry papers, 86 osteotomy registry papers and 157 cartilage registry papers (table 1). Our search identified 23 arthroplasty registries, 8 ligament registries, 4 osteotomy and 3 cartilage registries that fulfilled our inclusion criteria. Despite some inconsistencies in the most recent year of annual reports available, to date, the total numbers of primary knee replacements and revision/reoperation surgeries recorded in these registries are more than 4.451 million and 786 225, respectively (table 2).
Eight ligament registries yielded 154 830 records of primary ACLR surgeries and 14 754 records of reoperation following ACLR (table 3).
The cumulative number of recorded osteotomies in the available reports is over 7559. The type of osteotomy was available for the UKKOR and SKAR, of which 1184 were high tibial osteotomy procedures and 129 were distal femoral osteotomies (table 4). The cartilage registries document 28 744 records of primary cartilage repair/reconstruction surgeries (table 5).
Discussion
Registries have proven themselves as important tools for physicians, researchers, patients, industry and health providers.
One example of the impact of registries can be found in the study of Boyer et al.
They compared 190 registries-based publications with 476 from RCTs and 40 from meta-analyses. They found that registry-based studies had the highest impact between the three (calculating by the median number of citations in the 3-year period after publication).
The most prominent advantage of registries is their large cohort size, which in turn enables study of large populations and cases which are difficult to randomise,
Increased risk of ACL revision with non-surgical treatment of a concomitant medial collateral ligament injury: a study on 19,457 patients from the Swedish national knee ligament registry.
Furthermore, registries may affect perioperative standards of care as in the case of Jameson et al’s study that assessed the UK National Joint Registry (NJR), the Hospital Episode Statistics and the administrative database for all hospital admissions to the English National Health Service (NHS), and compared the aspirin with low-molecular-weight heparin (LMWH) for thromboprophylaxis in total knee arthroplasty (TKA) patients and reported no significant differences in terms of venous thromboembolism cases, 90-day mortality and major (cerebrovascular and gastrointestinal) haemorrhage. This demonstrated an association between aspirin use and a higher operative management for wound complications. Jameson et al’s group acknowledge their findings’ weaknesses including fundamental limitations of registry-based cohort such as lack of randomisation. For example, higher-risk patients were probably more likely to receive LMWH as a prophylactic measure. This in turn might result in underestimation of the effect of both treatments.
The effect of aspirin and low-molecular-weight heparin on venous thromboembolism after knee replacement: a non-randomised comparison using national joint registry data.
while RCTs’ conclusions are defined as level I evidence, and they are considered by many to be the most influential study design. RCTs have strict inclusion and exclusion criteria to compare an intervention against standard of care in equal groups. Registries provide information on the outcome of large numbers of patients with the real-world indications for surgery and patient results that extend beyond the confines of the narrow criteria of an RCT. By incorporating a well-planned framework to their protocols, one future role for clinical registries could be to serve as independent outcome instruments.
Another advantage of using registry data is the ability to compare the outcomes between registries and highlight pertinent differences found. Differences in patients’ characteristics, standard of care and clinical outcomes can give clinicians and decision makers another perspective to their findings and in some cases raise unexpected questions.
Over the years, registries have illustrated positive outcomes for surgical interventions studied, but also raised red flags when potentially harmful trends were recognised. Such was the case in the removal of the Articular Surface Replacement™ (ASR™) (DePuy Orthopaedics, Warsaw, Indiana, USA) following the negative outcomes, recognised in the 2007 UK NJR report.
While this report did not lead to a direct dismissal of the devices, it led to a regulatory authorities’ intervention and further evaluations which eventually lead to the desired outcome of poorly performing implants being removed from the marketplace.
Registry-based studies evaluate association, not causation. To exemplify this, Niinimäki et al who used the FNHDR (Finland) to evaluate the outcomes of high-tibial osteotomies measured by time of survival until conversion to TKA.
They reported an overall 10-year survivorship of 89%, with conversion to TKA as the endpoint. They reported a worse survivorship from 1998 to 2008 than in the years from 1987 to 1998. They concluded that one explanation for this difference was the lower experience of younger surgeons who may still have been demonstrating a learning curve effect due to the decreased in popularity as documented by the incidence of osteotomies and concomitant upward trend in the use of TKA.
This resulted in an effort to increase training of those learning osteotomy surgery and in keeping with UK guidelines for unicompartmental knee arthroplasty (UKA), potentially concentrating numbers in fewer centres, which would lessen the learning curve effect that they showed in their study and improve patients’ outcomes.
Another limitation of registry data is the inconsistency of data recorded and a lack of unification of terminology. One common difference which we found in the registries was the use of the term ‘revision’. While some registries address revision as any following surgery in which change of any component is replaced (the SKAR or the UK NJR), others differentiate reoperation and revision. Subsequent surgeries such as secondary patellar resurfacing are considered reoperation, and only removal and replacement of all the prosthesis are referred as revision in the Norwegian registry.
This may lead to a potential measurement bias which is the result of different data collection methods and assessment measurements. This often limits the validity of conclusions drawn from some registry-based studies. An example for this potential bias can be found in the UK NJR records in which secondary patellar resurfacing is analysed as a revision procedure.
However, the effectiveness of secondary patellar resurfacing is a matter of controversy with latest reports and estimation of 44%–64% satisfaction after this procedure,
with the additional risk of introducing potential surgery-related complications, all of which can affect the findings of success and complication rate following revision surgeries reported by the UK NJR.
In many arthroplasty registries, the main endpoint negative outcome measure was revision. However, with 20% of TKA patients reporting dissatisfaction, PROM measures have been added to registries as some failures do not reach revision surgery. Some have also added other measures such as records of complications and readmission. The use of TKA in younger patients follows different trends due to those patients often having increased demands and occasionally unrealistic expectations. This will require further attention and stratification of data. One problem with using PROMs is that some of them display a significant ceiling effect. This is a potential issue when the Oxford Knee Score or the American Knee Society Score is used.
Registries aim to record data of large number of cases. These large cohorts making the study of populations which are difficult to randomise, possible. Svantesson et al used the Swedish National Ligament Registry (NLR) to assess the long-term impact of concomitant medial and lateral collateral ligament (MCL and LCL, respectively) injuries on ACLR outcomes.
Increased risk of ACL revision with non-surgical treatment of a concomitant medial collateral ligament injury: a study on 19,457 patients from the Swedish national knee ligament registry.
Their goal was to compare the outcomes of ACLR surgeries with or without combined MCL/LCL treatment. Using this registry, they managed to compare 19 457 cases. Their results showed an increase risk of ACLR revision when concomitant MCL injury was not addressed at the time of the surgery. Interestingly, their study also reported less favourable patient reported outcomes at 2 years after the combined surgical treatment of anterior cruciate ligament (ACL) and MCL tears with better outcomes from conservative management of the MCL injury followed by ACLR. The presence of concomitant LCL injury did not seem to negatively affect the outcomes of ACLR surgeries compared with surgical treatment of isolated ACL tears.
One of the registries’ most fundamental advantages is their size. However, this can also become a potential limitation. Exceptionally large study populations can be used to statistically identify extremely subtle differences claiming ‘statistical significance’, which is commonly measured by small p values and narrow CIs. Statistical significance, however, does not always reflect a clinical significance and without careful interpretation may lead to incorrect decisions.
Another potential weakness of large datasets is that unless they report on the different outcomes of different models within an arthroplasty ‘family’, successful outcomes of one combination of such implants that is more commonly used may disguise poor outcomes from a different combination of models within the same family.
Since participation is voluntary in most registries, there is a potential selection bias. Moreover, costs and time requirements to assure high level of follow-up are high,
adding to the registries operative challenge. Encouraging enough, the UK experience of the NJR and the NLR reveals gradual increase in compliance since their establishments.
Potential selection bias can also be found when comparing different modalities of treatment, different techniques or different implants. This is not an exclusive pitfall of registry-based papers. Nevertheless, when assessing large numbers, it should be acknowledged when planning the methods for extracting data and analysing it. For example, when designing a comparative study between TKA and UKA, one must remember that UKA may be used in younger or older patients. Without proper planning and matching processes, analysis of these potentially two different populations may lead to skewed results.
Despite its many advantages, if clinicians and health authorities are not mandated to enter data, then uptake and consistency of data input can be an issue. Registries need governments or national health system funding in order to be successful on a health policy scale. Start-up costs are high for each registry and depending on the size of the measured countries and populations, costs may vary from several hundreds of thousands and to several millions of euros or US dollars.
A fledging registry needs the support of the relevant orthopaedic association and cooperation and commitment of their member clinicians. Without the involvement of health authorities and healthcare professionals, both establishment and maintenance of a fledgling registry are difficult to achieve.
Other costs include general maintenance and registration costs of implants, devices or operations. In order to be worth the investment, regulators and national health systems must recognise the value of establishing and operating registries. The importance of early recognition of adverse effects and early failures is one of the most important roles of registries. One key factor is the involvement of not only clinicians but also official regulators. Current medical devices reporting regulations of both the European Union and the USA aim to promote patients’ safety. Both require manufacturers to carry out postmarketing surveillance of new implants, evaluating benefits and harms of both newly introduced and well-established devices.
reviewed the Norwegian joint registry to compare the survivorship of hybrid and cemented TKAs. They found better or equal survivorship in the hybrid implants compared with the comparable cemented brand version laying to rest widely held concerns over this technique. They were able to show that in terms of time-efficiency hybrid, TKA can be a good option without increased risk of revision.
Another example of trend recognition can be found in the work of Niinimäki et al. They assessed the FNHDR and analysed the Finnish experience of knee osteotomies between the years 1987 and 2008. They witnessed a steady decrease in the number of osteotomies that were performed which correlated to the increased incidence of TKA.
However, a slight increase in the incidence of these procedures was seen in younger patients under 50 years of age. This emphasised that this technique is still a relevant surgical skill for OA treatment in younger patients.
Another advantage is the ability to compare between different registries. Bohm et al of the Canadian Joint Replacement Registry (CJRR) compared the CJRR to other main registries.
They identified a higher incidence of TKA in women in Canada compared with Sweden. Despite the suspected multifactorial causes for this difference, they hypothesised that this might also be the result of higher rates of obesity in Canada. Different trends in choice and outcomes of implant and fixation methods between the two national registry cohorts were also noted. Another example of such process was demonstrated by Prentice et al,
who compared data from six ACL registries. They reported similar demographics, injury characteristics, revision rates and incidence of surgical outcomes. However, several differences in patient management were observed, including time to reconstruction, concomitant injuries, graft selection and method of fixation.
Mohammad et al used the UK NJR’s data to compare cemented and cementless UKAs.
Comparison of the 10-year outcomes of cemented and cementless unicompartmental knee replacements: data from the National joint Registry for England, Wales, Northern Ireland and the Isle of man.
The first-generation UKA was cemented and associated with an increased 10-year revision rate compared with TKA. This and the more demanding technique led many surgeons to refrain from using it. However, over the years, with improvements in instrumentation and technique, UKA gained new attention. One of these developments was cementless fixation. However, randomised control studies and cohorts could not prove significant long-term differences between the two methods. One of the suggested reasons was the small number of patients participating in those studies. In order to overcome this issue, Mohammad et al, analysing the UK NJR’s data, showed that cementless UKA might be of merit as better 10-year implant survival rate was seen.
Comparison of the 10-year outcomes of cemented and cementless unicompartmental knee replacements: data from the National joint Registry for England, Wales, Northern Ireland and the Isle of man.
which found lower revision rate for cementless UKA. The main indication for revision in the cementless group was aseptic loosening. However, the cementless cohort was reported to have decreased unexplained postoperative pain and overall lysis; they suggested that this could be due to better fixation of the uncemented implants.
Adverse effects
Administrative claims data benefit from the high patient numbers, but their disadvantages include lack of precision of coding schemes.
When assessing complication rates following anterior cruciate ligament reconstruction in the English NHS 13 941 operations, they reported deep venous thrombosis (DVT) on 0.30% of the cases, pulmonary thromboembolism (PE) on 0.18%,
that the most common adverse outcomes were DVT 0.36%, effusion and synovitis 0.09%, PE 0.10%, haemarthrosis, readmission rate was 0.77%, and there were 55 deaths (0.03%). Overall, there were 6.4 adverse outcomes per 1000 elective knee arthroscopy procedures (0.64%).
Since 2015, there have been six annual reports from the UK NJR, none of which reported complication rates approaching those described previously. No complications are documented in the first report of the ICRS global registry. This was said to be typical of the earlier registry reports for soft tissue procedures, but in reality, complications will ultimately be shown to be in the frequencies reported for the arthroscopic techniques used for debridement and articular cartilage biopsy harvest and commonly associated procedures such as ACLR.
One example of postmarketing surveillance which led to a global intervention was the case of the market removal of ASR resurfacing system (DePuy Orthopaedics) following negative outcomes reported and recognised in the 2007 UK NJR report with a high 5-year revision rate.
These reports did not lead to a direct dismissal of the devices but rather lead to regulatory authorities’ intervention and further evaluations which eventually lead to the desired outcome of poorly performing implants being removed from the marketplace.
It has been stated that national arthroplasty registers monitoring of adverse outcomes are sufficiently robust that they ‘have to be used for research and post-market surveillance’.
Others disagree and conclude that these data may be too insensitive, such as using revision as an endpoint for the success of knee arthroplasty as it was exaggerating the success of total knee replacements in comparison to partial knee replacement.
Several factors may play a role here such as inconsistently attributed causes of failure between registries or patients’ inaccurate reporting of the presence of complications but not their absence. The sensitivity of patient-reported complications, for example, needs to be improved.
The information registries generated may complement more traditional data collection systems employed in postmarket surveillance studies but is not designed to replace them.
Cost-effectiveness assessment
With appropriate design of data collection, registries may be able to assess cost-effectiveness. Weeks et al
used the Australian National Joint Replacement Registry in order to assess and compare the cost effectiveness of TKA with or without patellar resurfacing. After analysing the 2014 annual report of the Australian registry, assessing the outcome measures and estimating the costs from a Canadian point of view, their country of interest, they reported lower revision rates for TKA with patellar resurfacing. With a 14-year review of a retrospective cohort, they reported lower costs ($12 917.01 vs $13 296.63) and higher quality-adjusted life year. This, in turn, led them to conclude that patellar resurfacing in TKA is cost-effective. Their results are important for two main reasons. First, counterintuitively, they found that despite a longer operating time and the cost of the patellar implant, this more time-consuming procedure was worth performing. Their second and more important implication was their use of another country’s national registry in order to answer a question which was relevant to their own health system policy.
Registries can help predict change in demands of different populations
The epidemiology of revision total knee and hip arthroplasty in England and Wales: a comparative analysis with projections for the United States. A study using the National joint registry dataset.
and, in turn, could be used for policy making when health administrators plan the costs and future needs of a health system.
In their 2015 comparative analysis of the NJR, Patel et al predicted a major increase in demand for both primary TKA and revision TKA from 2012 to 2030. Despite some limitations which relate to general features of the NJR and other registries, their projections had implications for the economic burden which they were expecting to witness.
The epidemiology of revision total knee and hip arthroplasty in England and Wales: a comparative analysis with projections for the United States. A study using the National joint registry dataset.
They also compared their estimations to those of other registries such as the Swedish and the New Zealand’s arthroplasty registries and found similar trends. These examples emphasise the importance to healthcare administrators and decision makers with involvement and use of registries.
Conclusion
Registries have become important instruments for physicians, researchers, patients, health economists, industry and health providers when selecting patient treatment and designing health policies. Over the past three decades, registries have evolved significantly. The efficiency of collection and the coverage continue to improve. Nevertheless, even with the improvements we have seen over the years, it is increasingly recognised that collection of high-quality and surgeon-validated data using validated scoring systems which are sensitive to change and do not have ceiling effects is of paramount importance.
In order to meet these needs, cooperation of governments, professional associations and clinicians is crucial for funding, operating and assessing these registries. Robust and accurate adverse event documentation is required in order to convince regulatory authorities of the safety of knee procedures under their review. Knee registries are adding useful information to aid decision making for all interested parties, but that there is a constant evolution ongoing which will continue to improve the ease of data input, allowing greater understanding of such generalisable real-world results.
Kolling C, Simmen BR, Labek G, Goldhahn J. Key factors for a successful National Arthroplasty Register. J Bone Joint Surg Br. 2007 Dec;89(12):1567–73.
•
Boyer P, Boutron I, Ravaud P. Scientific production and impact of national registers: the example of orthopaedic national registers.Osteoarthritis and Cartilage. 2011 Jul;19(7):858–63.
•
Rolfson O, Bohm E, Franklin P, Lyman S, Denissen G, Dawson J, et al. Patient-reported outcome measures in arthroplasty registries: Report of the Patient-Reported Outcome Measures Working Group of the International Society of Arthroplasty RegistriesPart II. Recommendations for selection, administration, and analysis. Acta Orthopaedica. 2016 Jun 15;87(sup1):9–23.
•
Lübbeke A, Silman AJ, Barea C, Prieto-Alhambra D, Carr AJ. Mapping existing hip and knee replacement registries in Europe. Health Policy. 2018 May;122(5):548–57.
•
Wilson I, Bohm E, Lübbeke A, Lyman S, Overgaard S, Rolfson O, et al. Orthopaedic registries with patient-reported outcome measures. EFORT Open Reviews. 2019 Jun;4 (6):357–67.
•
Biant LC, Conley CW, McNicholas MJ. The First Report of the International Cartilage Regeneration and Joint Preservation Society’s Global Registry. Cartilage. 2020 Feb 19;1947603520907664.
Increased involvement of health systems, professional associations and clinicians for funding, operating and assessing these registries.
•
Global collaboration between registries.
•
Registries as a more accessible data source.
Acknowledgements
EB-N has a research Fellowship in Manchester University NHS foundation trust, under the EFORT scholarship supported by Medacta International.
Contributors
All authors certify that they have participated sufficiently in the work to take public responsibility for the content. EB and MJM conceptualised and designed the work, participated in the acquisition and analysis of the data. All authors participated in interpretation of the data. EB and MJM drafted the manuscript together with NN and CL who helped editing it critically for important intellectual content. All authors gave final approval of the version to be published and agree to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.
Funding
The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.
Competing interests
MJMcN is Chair of the ICRS Global Registry Committee and on the Steering Committee of the UK National Ligament Registry. NN is a Journal of ISAKOS Editorial Board member.
Provenance and peer review
Commissioned; externally peer reviewed.
Ethics statements
Patient consent for publication Not required.
References
Lübbeke A.
Silman A.J.
Barea C.
et al.
Mapping existing hip and knee replacement registries in Europe.
The relationship between the angle of the tibial tunnel in the coronal plane and loss of flexion and anterior laxity after anterior cruciate ligament reconstruction.
Degenerative isolated cartilage defects of the patellofemoral joint are associated with more severe symptoms compared to trauma-related defects: results of the German cartilage registry (KnorpelRegister DGOU).
Increased risk of ACL revision with non-surgical treatment of a concomitant medial collateral ligament injury: a study on 19,457 patients from the Swedish national knee ligament registry.
The effect of aspirin and low-molecular-weight heparin on venous thromboembolism after knee replacement: a non-randomised comparison using national joint registry data.
Comparison of the 10-year outcomes of cemented and cementless unicompartmental knee replacements: data from the National joint Registry for England, Wales, Northern Ireland and the Isle of man.
The epidemiology of revision total knee and hip arthroplasty in England and Wales: a comparative analysis with projections for the United States. A study using the National joint registry dataset.
00306-0&id=gr1.jpg){kind=link}