Search

Registries have significant roles in assessing the comparative performance of devices. Ideally, early identification of outliers should use a time-to-event outcome while reducing the confounding effects of other components in the device and patient characteristics. Machine learning (ML), which contains self-learning algorithms, is one approach to consider many variables simultaneously to reduce the impact of confounding. The principal objective of this study was to investigate the effectiveness of using either random survival forest (RSF) or regularised/unregularised Cox regression to account for patient and associated device confounding factors in comparison with current standard techniques. This study evaluated RSF and regularised/unregularised Cox regression using data from the Australian Orthopaedic Association National Joint Replacement Registry (AOANJRR) to detect outlier devices among 213 individual primary total hip components performed in 163,356 primary procedures from 1 January 2015 to the end of 2019. Device components and patient characteristics were the inputs, and time to first revision surgery was the primary outcome treated as a censored case for death. The effectiveness of the ML approaches was assessed based on the ability to detect the outliers identified by the AOANJRR standard approach. In the study cohort, the standardised AOANJRR approach identified three acetabular components and seven femoral stems as outliers. The ML approaches identified some but not all the outliers detected by the AOANJRR. Both the methods identified three of the same femoral stems, and the RSF identified the other five components, including two of the same acetabular cups and three of the same femoral stems. In addition, both the RSF and Cox techniques detected a number of additional device components that were not previously identified by the standard approach. The results showed that ML may be able to offer a supplementary approach to enhance the early identification of outlier devices. Random survival forest was a more comparable technique to the AOANJRR standard than the Cox regression, but further studies are required to better understand the potential of ML to improve the early identification of outliers. [ABSTRACT FROM AUTHOR]

Prediction models are being increasingly used in the medical field to identify risk factors and possible outcomes. Some of these are presently being used to develop guidelines for improving clinical practice. The application of machine learning (ML), comprising a powerful set of computational tools for analysing data, has been clearly expanding in the role of predictive modelling. This paper reviews the latest developments of supervised ML techniques that have been used to analyse data related to post‐operative total hip and knee replacements. The aim was to review the most recent findings of relevant published studies by outlining the methodologies employed (most‐widely used supervised ML techniques), data sources, domains, limitations of predictive analytics and the quality of predictions. This paper reviews the latest developments of supervised machine learning (ML) techniques that have been used to analyse data related to post‐operative total hip and knee replacements. The aim was to review the most recent findings of relevant published studies by outlining the methodologies employed (most widely used supervised ML techniques), data sources, domains, limitations of predictive analytics and the quality of predictions. [ABSTRACT FROM AUTHOR]

Staphylococcus aureus is a major causative pathogen of osteomyelitis. Intracellular infections of resident bone cells including osteocytes can persist despite goldstandard clinical intervention. The mechanisms by which intracellular S. aureus evades antibiotic therapy are unknown. In this study, we utilised an in vitro S. aureus infection model of human osteocytes to investigate whether antibioticmediated dysregulation of autophagy contributes to this phenomenon. Infected or non-infected osteocyte-like cells were exposed to combinations of rifampicin, vancomycin, and modulators of autophagy. Intracellular bacterial growth characteristics were assessed using colony-forming unit (CFU) analysis, viable bacterial DNA abundance, and the rate of escape into antibiotic-free medium, together with measures of autophagic flux. Rifampicin, alone or in combination with vancomycin, caused a rapid decrease in the culturability of intracellular bacteria, concomitant with stable or increased absolute bacterial DNA levels. Both antibiotics significantly inhibited autophagic flux. However, modulation of autophagic flux did not affect viable bacterial DNA levels. In summary, autophagy was shown to be a factor in the host--pathogen relationship in this model, as itsmodulation affected the growth state of intracellular S. aureus with respect to both their culturability and propensity to escape the intracellular niche. While rifampicin and vancomycin treatments moderately suppressed autophagic flux acutely, this did not explain the paradoxical response of antibiotic treatment in decreasing S. aureus culturability whilst failing to clear bacterial DNA and hence intracellular bacterial load. Thus, offtarget effects of rifampicin and vancomycin on autophagic flux in osteocyte-like cells could not explain the persistent S. aureus infection in these cells. [ABSTRACT FROM AUTHOR]

Background: There are variations in the performance of individual prostheses used in hip replacements. Some of which have unexpectedly higher revision rates – outliers. The Australian Orthopaedic Association National Joint Replacement Registry (AOANJRR) has established a standardised multi-stage approach for identifying these devices. This is done by comparing the revision rates of individual prostheses to all other prostheses in class, with the exception of large head metal-on-metal (LHMoM) prostheses. However, improvements in device design and performance over time have required a need to reconsider the comparator group. This study aimed to identify a more specific comparator to better reflect contemporary surgical practice. Methods: The time to first revision was estimated on the data of 413,417 primary total conventional hip replacements undertaken for osteoarthritis (OA) from 01 January 2003 to 31 December 2019. Survivorship analyses with stepwise exclusions were undertaken. The first exclusion was LHMoM, followed by other non-modern bearing surfaces (defined as all the bearing couples except metal or ceramic heads on cross-linked polyethylene and mixed ceramic-on-ceramic), and then devices with modular neck-stem design or used for specific purposes (incl. constrained, dual-mobility, and head size <28 mm). Lastly, all remaining prostheses previously identified as having a higher than anticipated rate of revision (HTARR) were also excluded. Results: These exclusions progressively reduced the cumulative percent revision (CPR) rate. The final comparator, which only includes satisfactory-performed prostheses of contemporary design and use, has a 10-year CPR of 4.30% (95% CI, 4.2–4.41) which is lower than 4.93% (95% CI, 4.84–5.02) for the current comparator used by the AOANJRR (all prostheses excluding LHMOM). Over the study period, 13 additional components were identified utilising the modified comparator. Conclusions: The calculation of the comparator revision rate should be re-evaluated to include only modern prosthesis constructs to ensure that poorly performing prostheses are identified early. [ABSTRACT FROM AUTHOR]

While finite element (FE) models have been used extensively in orthopedic studies, validation of their outcome metrics has been limited to comparison against ex vivo testing. The aim of this study was to validate FE model predictions of the initial cup mechanical environment against patient‐matched in vivo measurements of acetabular cup migration using radiostereometric analysis (RSA). Tailored musculoskeletal and FE models were developed using a combination of three‐dimensional (3D) motion capture data and clinical computerized tomography (CT) scans for a cohort of eight individuals who underwent primary total hip replacement and were prospectively enrolled in an RSA study. FE models were developed to calculate the mean modulus of cancellous bone, composite peak micromotion (CPM), composite peak strain (CPS) and percentage area of bone ingrowth. The RSA cup migration at 3 months was used to corroborate the FE output metrics. Qualitatively, all FE‐predicted metrics followed a similar rank order as the in vivo RSA 3D migration data. The two cases with the lowest predicted CPM (<20 µm), lowest CPS (<0.0041), and high bone modulus (>917 MPa) were confirmed to have the lowest in vivo RSA 3D migration (<0.14 mm). The two cases with the largest predicted CPM (>80 µm), larger CPS (>0.0119) and lowest bone modulus (<472 MPa) were confirmed to have the largest in vivo RSA 3D migration (>0.78 mm). This study enabled the first corroboration between tailored musculoskeletal and FE model predictions with in vivo RSA cup migration. Investigation of additional patient‐matched CT, gait, and RSA examinations may allow further development and validation of FE models. [ABSTRACT FROM AUTHOR]

The intracellular infection of osteocytes represents a clinically important aspect of osteomyelitis. However, few human osteocyte in vitro models exist and the differentiation of immature osteoblasts to an osteocyte stage typically takes at least 4‐weeks of culture, making the study of this process challenging and time consuming. The osteosarcoma cell line Saos‐2 has proved to be a useful model of human osteoblast to mature osteocyte differentiation. Culture under osteogenic conditions in a standard normoxic (21% O2) atmosphere results in reproducible mineralization and acquisition of mature osteocyte markers over the expected 28–35 day culture period. In order to expedite experimental assays, we tested whether reducing available oxygen to mimic concentrations experienced by osteocytes in vivo would increase the rate of differentiation. Cells cultured under 1% O2 exhibited maximal mineral deposition by 14 days. Early (COLA1, MEPE) and mature (PHEX, DMP1, GJA1, SOST) osteocyte markers were upregulated earlier under hypoxia compared to normoxia. Cells differentiated under 1% O2 for 14 days displayed a similar ability to internalize Staphylococcus aureus as day 28 cells grown under normoxic conditions. Thus, low oxygen accelerates Saos‐2 osteocyte differentiation, resulting in a useful human osteocyte‐like cell model within 14 days. [ABSTRACT FROM AUTHOR]

Fretting corrosion is a known failure mechanism of total hip replacement (THR) that can lead to revision surgery. Implant retrieval studies have thoroughly documented the occurrence of fretting corrosion in THR implants and its correlation with implant- and patient-related factors. Although implant retrieval studies benefit both clinicians and implant manufacturers, the limitations of these types of studies need to be acknowledged. For example, while some factors are routinely investigated for a possible correlation with failure due to fretting corrosion, other factors are often assumed to have no effect. To improve on these limitations, this review investigates the most significant patient- and implant-related risk factors for fretting corrosion of THR implants for both published retrieval studies and joint replacement registries. The findings and limitations are discussed critically. It is concluded that retrieval studies add significant insight into implant failure mechanisms and should be used in conjunction with joint replacement registry reports. It is suggested that the development of reliable predictive models based on implant failure risk factors and decision-making support systems could lead to enhanced implant longevity. [ABSTRACT FROM AUTHOR]

The aim of this study was to determine the effect of surgical change to the acetabular offset and femoral offset on the abductor muscle and hip contact forces after primary total hip arthroplasty (THA) using computational methods. Thirty‐five patients undergoing primary THA were recruited. Patients underwent a computed tomography scan of their pelvis and hip, and underwent gait analysis pre‐ and 6‐months postoperatively. Surgically induced changes in acetabular and femoral offset were used to inform a musculoskeletal model to estimated abductor muscle and hip joint contact forces. Two experiments were performed: (1) influence of changes in hip geometry on hip biomechanics with preoperative kinematics; and (2) influence of changes in hip geometry on hip biomechanics with postoperative kinematics. Superior and medial placement of the hip centre of rotation during THA was most influential in reducing hip contact forces, predicting 63% of the variance (p < 0.001). When comparing the preoperative geometry and kinematics model, with postoperative geometry and kinematics, hip contact forces increased after surgery (0.68 BW, p = 0.001). Increasing the abductor lever arm reduced abductor muscle force by 28% (p < 0.001) and resultant hip contact force by 17% (0.6 BW, p = 0.003), with both preoperative and postoperative kinematics. Failure to increase abductor lever arm increased resultant hip contact force 11% (0.33 BW, p < 0.001). In conclusion, increasing the abductor lever arm provides a substantial biomechanical benefit to reduce hip abductor and resultant hip joint contact forces. The magnitude of this effect is equivalent to the average increase in hip contact force seen with improved gait from pre‐to post‐surgery. [ABSTRACT FROM AUTHOR]

Metal artifact reduction (MAR) has improved computed tomography (CT) imaging of total hip arthroplasty (THA) but the assessment of osteolysis and implant to bone contact relies on the accurate depiction of bone defects, cancellous bone, and cement. This study evaluates the impact of available single and dual‐energy protocols on periprosthetic tissue characterization in a cadaveric phantom. Bilateral THA was performed on a fresh frozen cadaveric pelvis with simulated osteolytic cavities. CT acquisitions with projection‐based MAR and noise equivalence were performed using single energy 140 kVp, single energy 150 kVp with 0.6 mm tin filtration, and dual‐energy at 100/150 kVp with 0.6 mm tin filtration, from which simulated energies were extracted. Image subtraction, segmentation, region of interest histograms, and line profiles were used to characterize tissue density and separation. Tissue densities were heavily dependent on the energy profile of the protocol. Cancellous bone ranged from 182 to 45 HU and cement from 1012 to 131 HU using 140 kVp compared to dual‐energy with weighted high energy tube, respectively. Spectral separation between cancellous bone, osteolytic defect, and cement was reduced for all protocols compared with 140 kVp. Spectral overlap was most severe using dual‐energy with heavily weighted high‐energy tubes. Dual‐energy algorithms reduced trabecular contrast within the cancellous bone and cortical edge response. Although the dual‐energy acquisition has been proposed as an additive to projection‐based MAR techniques in THA, reduced density and contrast in clinically relevant periprosthetic tissue compared to 140 kVp single energy may limit its use in characterizing periprosthetic tissues. [ABSTRACT FROM AUTHOR]

Background and purpose -- We conducted a systematic review and meta-analysis of RSA studies to investigate the early and long-term migration patterns of acetabular cups and the influence of implant factors on cup migration over time. Methods -- We performed a systematic search of PubMed, Embase, and Scopus databases to identify all RSA studies of cup migration following primary total hip replacement (THR). Proximal migration at 3 and 6 months, 1, 2, 5, and 10 years were considered for analysis. Implant factors investigated included fixation type, head size, bearing surface, uncemented coating design, and the decade of RSA introduction. Results -- 47 studies reported the proximal migration of 83 cohorts (2,328 cups). Besides 1 threaded cup design, no implant factor investigated was found to significantly influence proximal migration. The mean pooled 2-year proximal migration of cemented cups (0.14 mm, 95% confidence interval [CI] 0.08-0.20) was not significantly different from uncemented cups (0.12 mm, CI 0.04-0.19). The mean pooled proximal migration at 6 months was 0.11 mm (CI 0.06-0.16) and there was no significant increase between 6 months and 2 years (0.015 mm, CI 0.000-0.030). 27 of 75 cohorts (36%) reported mean proximal migration greater than 0.2 mm at 2 years, which has previously been identified as a predictor of implants at risk of long-term loosening. Conclusion -- Our meta-analysis demonstrated that the majority of cup migration occurs within the first 6 months. With one exception, no implant factors influenced the 2-year proximal migration of acetabular cups. 36% of studies with 2-year migration were considered at risk of long-term loosening. Further investigation and comparison against longterm survivorship data would validate 6-month and/or 1-year proximal migration measurements as an earlier predictor of long-term loosening than the current 2-year threshold. [ABSTRACT FROM AUTHOR]

Polyethylene (PE) liners are a common bearing surface of orthopaedic prostheses. Wear particles of ultra-high molecular weight PE (UHMWPE) contribute to periprosthetic osteolysis, a major cause of aseptic loosening. Vitamin E is added to some PE liners to prevent oxidative degradation. Osteocytes, an important cell type for controlling both bone mineralisation and bone resorption, have been shown to respond UHMWPE particles by upregulating pro-osteoclastogenic and osteocytic osteolysis. Here, we examined the effects of the vitamin E analogues α-tocopherol and γ-tocotrienol alone or in the context of UHMWPE particles on human osteocyte gene expression and mineralisation behaviour. Human osteoblasts differentiated to an osteocyte-like stage were exposed to UHMWPE wear particles in the presence or absence of either α-Tocopherol or γ-Tocotrienol. Both α-Tocopherol and γ-Tocotrienol induced antioxidant-related gene expression. UHMWPE particles independently upregulated antioxidant gene expression, suggesting an effect of wear particles on oxidative stress. Both vitamin E analogues strongly induced OPG mRNA expression and γ-Tocotrienol also inhibited RANKL mRNA expression, resulting in a significantly reduced RANKL:OPG mRNA ratio (p < 0.01) overall. UHMWPE particles reversed the suppressive effect of α-Tocopherol but not of γ-Tocotrienol on this pro-osteoclastogenic index. UHMWPE particles also upregulated osteocytic-osteolysis related gene expression. Vitamin E analogues alone or in combination with UHMWPE particles also resulted in upregulation of these genes. Consistent with this, both vitamin E analogues promoted calcium release from mineralised cultures of osteocyte-like cells. Our findings suggest that while vitamin E may suppress osteocyte support of osteoclastogenesis in the presence of UHMWPE particles, the antioxidant effect may induce osteocytic osteolysis, which could promote periprosthetic osteolysis. It will be important to conduct further studies of vitamin E to determine the long-term effects of its inclusion in prosthetic materials. [ABSTRACT FROM AUTHOR]

Biomechanical factors (e.g., joint loading) have a significant role in the progression of osteoarthritis (OA). However, some relationships between in vivo joint loading indices and tibial cartilage thickness are conflicting. This study investigated relationships between pre‐operative in vivo external knee joint moments, joint alignment and regional tibial cartilage thickness using micro‐CT in subjects with end‐stage knee OA. Tibial plateaus from 25 patients that underwent knee replacement for OA were micro‐CT scanned (17 µm/voxel). Prior to surgery, subjects underwent gait analysis to calculate external knee moments. The mechanical axis deviation (MAD) was obtained from pre‐operative radiographs. Cartilage thickness (Cart.Th) was analyzed from micro‐CT images, in anteromedial, anterolateral, posteromedial and posterolateral subregions of interest. Medial‐to‐lateral Cart.Th ratios were also explored. Relationships between Cart.Th and joint loading indices were examined using Pearson's correlations. Significant correlations were found between Cart.Th and joint loading indices, positive anteromedially with the first peak knee adduction moment (r = 0.55, p < 0.01) and external rotation moment (ERM; r = 0.52, p < 0.01), and negative with MAD (r = −0.76, p < 0.001). In the lateral regions, these correlations had opposite signs. The medial‐to‐lateral Cart.Th ratio correlated strongly with ERM (r = 0.63, p = 0.001) and MAD (r = −0.75, p < 0.001). Joint loading indices correlated with regional cartilage thickness values and their medial‐to‐lateral ratios in end‐stage knee OA subjects, with higher regional loads corresponding to thinner cartilage. These relationships have the opposite sign compared to the subchondral bone microarchitecture found in our previous study on the same specimens, which may suggest a complementary bone–cartilage interplay in response to loading. [ABSTRACT FROM AUTHOR]

The objectives of this study were to (1) develop a semiautomated method to obtain lesion volume and bone mineral density (BMD) in terms of Hounsfield units from pelvic computed tomography (CT) scans in three regions of interest, and (2) assess accuracy and reliability of the method based on cadaveric CT scans. Image artefacts due to metal implants reduce CT clarity and are more severe with more than one implant in situ. Therefore, accuracy and reliability tests were performed with varying numbers of total hip arthroplasties implanted. To test the accuracy of lesion size measurements, microcomputed tomography was used as a reference. Mean absolute error ranged from 36 to 284 mm3 after five measurements. Intra‐ and inter‐operator reliability of the entire method was measured for a selection of parameters. All coefficient of variation values were good to excellent for CT scans of the native pelvic anatomy and a CT scans of the same pelvis with one and two implants in situ. Accuracy of quantifying lesion volume decreased with decreasing CT image clarity by 0.6%–3.6% mean absolute relative error. Reliability of lesion volume measurement decreased with decreasing CT clarity. This was also the case for reliability of BMD measurements in the region most disrupted by metal artefact. The presented method proposes an approach for quantifying bone loss which has been proven to be accurate, reliable, and clinically applicable. [ABSTRACT FROM AUTHOR]

Infectious osteomyelitis associated with periprosthetic joint infections is often recalcitrant to treatment and has a high rate of recurrence. In the case of Staphylococcus aureus , the most common pathogen in all forms of osteomyelitis, this may be attributed in part to residual intracellular infection of host cells, yet this is not generally considered in the treatment strategy. Osteocytes represent a unique cell type in this context due to their abundance, their formation of a syncytium throughout the bone that could facilitate bacterial spread and their relative inaccessibility to professional immune cells. As such, there is potential value in studying the host-pathogen interactions in the context of this cell type in a replicable and scalable in vitro model. Here, we examined the utility of the human osteosarcoma cell line SaOS2 differentiated to an osteocyte-like stage (SaOS2-OY) as an intracellular infection model for S. aureus. We demonstrate that S. aureus is capable of generating stable intracellular infections in SaOS2-OY cells but not in undifferentiated, osteoblast-like SaOS2 cells (SaOS2-OB). In SaOS2-OY cells, S. aureus transitioned towards a quasi-dormant small colony variant (SCV) growth phenotype over a 15-day post-infection period. The infected cells exhibited changes in the expression of key immunomodulatory mediators that are consistent with the infection response of primary osteocytes. Thus, SaOS2-OY is an appropriate cell line model that may be predictive of the interactions between S. aureus and human osteocytes, and this will be useful for studying mechanisms of persistence and for testing the efficacy of potential antimicrobial strategies. [ABSTRACT FROM AUTHOR]

Cognitive decline and osteoporosis often coexist and some evidence suggests a causal link. However, there are no data on the longitudinal relationship between cognitive decline, bone loss and fracture risk, independent of aging. This study aimed to determine the association between: (i) cognitive decline and bone loss; and (ii) clinically significant cognitive decline (≥3 points) on Mini Mental State Examination (MMSE) over the first 5 years and subsequent fracture risk over the following 10 years. A total of 1741 women and 620 men aged ≥65 years from the population‐based Canadian Multicentre Osteoporosis Study were followed from 1997 to 2013. Association between cognitive decline and (i) bone loss was estimated using mixed‐effects models; and (ii) fracture risk was estimated using adjusted Cox models. Over 95% of participants had normal cognition at baseline (MMSE ≥ 24). The annual % change in MMSE was similar for both genders (women −0.33, interquartile range [IQR] −0.70 to +0.00; and men −0.34, IQR: −0.99 to 0.01). After multivariable adjustment, cognitive decline was associated with bone loss in women (6.5%; 95% confidence interval [CI], 3.2% to 9.9% for each percent decline in MMSE from baseline) but not men. Approximately 13% of participants experienced significant cognitive decline by year 5. In women, fracture risk was increased significantly (multivariable hazard ratio [HR], 1.61; 95% CI, 1.11 to 2.34). There were too few men to analyze. There was a significant association between cognitive decline and both bone loss and fracture risk, independent of aging, in women. Further studies are needed to determine mechanisms that link these common conditions. © 2021 American Society for Bone and Mineral Research (ASBMR). [ABSTRACT FROM AUTHOR]