Successful estimation of postoperative PROMs prior to a joint replacement surgery is important in deciding the best treatment option for a patient. However, estimation of the outcome is associated with substantial noise around individual prediction. Here, we test whether a classifier neural network can be used to simultaneously estimate postoperative PROMs and uncertainty better than current methods. We perform Oxford hip score (OHS) estimation using data collected by the NJR from 249,634 hip replacement surgeries performed from 2009 to 2018. The root mean square error (RMSE) of the various methods are compared to the standard deviation of outcome change distribution to measure the proportion of the total outcome variability that the model can capture. The area under the curve (AUC) for the probability of the change score being above a certain threshold was also plotted. The proposed classifier NN had a better or equivalent RMSE than all other currently used models. The threshold AUC shows similar results for all methods close to a change score of 20 but demonstrates better accuracy of the classifier neural network close to 0 change and greater than 30 change, showing that the full probability distribution performed by the classifier neural network resulted in a significant improvement in estimating the upper and lower quantiles of the change score probability distribution. Consequently, probabilistic estimation as performed by the classifier NN is the most adequate approach to this problem, since the final score has an important component of uncertainty. This study shows the importance of uncertainty estimation to accompany postoperative PROMs prediction and presents a clinically-meaningful method for personalised outcome that includes such uncertainty estimation.
Increased collection of patient-reported outcome measures (PROM) in registries enables international comparison of patient-centered outcomes after knee and hip replacement. We aimed to investigate 1) variations in PROM improvement, 2) the possible confounding factor of BMI, and 3) differences in comorbidity distributions between registries. Registries affiliated with the International Society of Arthroplasty Registries (ISAR) or OECD membership countries were invited to report aggregate EQ-5D, OKS, OHS, HOOS-PS and KOOS-PS values. Eligible patients underwent primary total, unilateral knee or hip replacement for osteoarthritis within three years and had completed PROMs preoperatively and either 6 or 12 months postoperatively, excluding patients with subsequent revisions. For each PROM cohort, Chi-square tests were performed for BMI distributions across registries and 12 predefined PROM strata (male/female, age 20-64/65-74/>75, high or low preoperative PROM scores). Comorbidity distributions were reported for available comorbidity indexes. Thirteen registries from 9 countries contributed data, n~130000 knee (range 140 to 79848) and n~113000 hip (range 137 to 85281). Mean EQ-5D index values (10 registries) ranged from 0.53 to 0.71 (knee) and 0.50 to 0.70 (hips) preoperatively and 0.78 to 0.85 (knee) and 0.83 to 0.87 (hip) postoperatively. Mean OKS (6 registries) ranged from 19.3 to 23.6 preoperatively and 36.2 to 41.2 postoperatively. Mean OHS (7 registries) ranged from 18.0 to 23.2 preoperatively and 39.8 to 44.2 postoperatively. Four registries reported KOOS-PS and three reported HOOS-PS. Proportions of patients with BMI >30 ranged from 35 to 62% (10 knee registries) and 16 to 43% (11 hip registries). For both knee and hip registries, distributions of patients across six BMI categories differed significantly among registries (p30 were for patients in the youngest age groups (20 to 64 and 65 to 74 years) with the lowest baseline scores. Additionally, females with lowest preoperative PROM scores had highest BMI. These findings were echoed for the OHS and OKS cohorts. Proportions of patients with ASA scores ≥3 ranged from 7 to 42% (9 knee registries) and 6 to 35% (8 hip registries). PROM-score improvement varies between international registries, which may be partially explained by differences in age, sex and preoperative scores. BMI and comorbidity may be relevant to adjust for.
Data of high quality are critical for the meaningful interpretation of registry information. The National Joint Registry (NJR) was established in 2002 as the result of an unexpectedly high failure rate of a cemented total hip arthroplasty. The NJR began data collection in 2003. In this study we report on the outcomes following the establishment of a formal data quality (DQ) audit process within the NJR, within which each patient episode entry is validated against the hospital unit’s Patient Administration System and vice-versa. This process enables bidirectional validation of every NJR entry and retrospective correction of any errors in the dataset. In 2014/15 baseline average compliance was 92.6% and this increased year-on-year with repeated audit cycles to 96.0% in 2018/19, with 76.4% of units achieving > 95% compliance. Following the closure of the audit cycle, an overall compliance rate of 97.9% was achieved for the 2018/19 period. An automated system was initiated in 2018 to reduce administrative burden and to integrate the DQ process into standard workflows. Our processes and quality improvement results demonstrate that DQ may be implemented successfully at national level, while minimizing the burden on hospitals. Cite this article:
Our main aim was to describe the trend in the comorbidities of patients undergoing elective total hip arthroplasties (THAs) and knee arthroplasties (KAs) between 1 January 2005 and 31 December 2018 in England. We combined data from the National Joint Registry (NJR) on primary elective hip and knee arthroplasties performed between 2005 and 2018 with pre-existing conditions recorded at the time of their primary operation from Hospital Episodes Statistics. We described the temporal trend in the number of comorbidities identified using the Charlson Comorbidity Index, and how this varied by age, sex, American Society of Anesthesiologists (ASA) grade, index of multiple deprivation, and type of KA.Aims
Methods
A recent report from France suggested an association between the use of cobalt-chrome (CoCr) femoral heads in total hip arthroplasties (THAs) and an increased risk of dilated cardiomyopathy and heart failure. CoCr is a commonly used material in orthopaedic implants. If the reported association is causal, the consequences would be significant given the millions of joint arthroplasties and other orthopaedic procedures in which CoCr is used annually. We examined whether CoCr-containing THAs were associated with an increased risk of all-cause mortality, heart outcomes, cancer, and neurodegenerative disorders in a large national database. Data from the National Joint Registry was linked to NHS English hospital inpatient episodes for 374,359 primary THAs with up to 14.5 years' follow-up. We excluded any patients with bilateral THAs, knee arthroplasties, indications other than osteoarthritis, aged under 55 years, and diagnosis of one or more outcome of interest before THA. Implants were grouped as either containing CoCr or not containing CoCr. The association between implant construct and the risk of all-cause mortality and incident heart failure, cancer, and neurodegenerative disorders was examined.Aims
Methods
The term heterotopic ossification (HO) describes lamellar bone formation within soft tissues following injury. A genome-wide scan of patients after hip arthroplasty has identified that variation within the lncRNA CASC20 is associated with HO susceptibility. Previous findings in our lab have demonstrated upregulation of CASC20 during BMP2-induced osteodifferentiation of adipose-derived stem cells (hMAD) alongside osteodifferentiation markers, RUNX2 and OSX. We hypothesize that CASC20 is a novel regulator of bone formation and aim to investigate CASC20 function in bone formation. 1) We used miRanda prediction algorithm and the ENCORI database to respectively predict which miRNAs CASC20 interacts with and to select for experimentally validated miRNAs. 2) We characterized the expression and functional role of CASC20-interacting miRNAs by respectively analyzing publicly available datasets (GSE107279 and pubmed.ncbi.nlm.nih.gov/26175215/) and by using Gene Ontology (GO) analysis. 3) We overexpressed CASC20 in hMAD using a lentiviral system and tested the effect of CASC20 overexpression in osteodifferentiation and expression of putative CASC20-interacting miRNAs.Abstract
Objectives
Methods
Conventional approaches (including Tobit) do not accurately account for ceiling effects in PROMs nor give uncertainty estimates. Here, a classifier neural network was used to estimate postoperative PROMs prior to surgery and compared with conventional methods. The Oxford Knee Score (OKS) and the Oxford Hip Score (OHS) were estimated with separate models. English NJR data from 2009 to 2018 was used, with 278.655 knee and 249.634 hip replacements. For both OKS and OHS estimations, the input variables included age, BMI, surgery date, sex, ASA, thromboprophylaxis, anaesthetic and preoperative PROMs responses. Bearing, fixation, head size and approach were also included for OHS and knee type for OKS estimation. A classifier neural network (NN) was compared with linear or Tobit regression, XGB and regression NN. The performance metrics were the root mean square error (RMSE), maximum absolute error (MAE) and area under curve (AUC). 95% confidence intervals were computed using 5-fold cross-validation.Abstract
Objectives
Methods
Debate remains whether the patella should be resurfaced during total knee replacement (TKR). For non-resurfaced TKRs, we estimated what the revision rate would have been if the patella had been resurfaced, and examined the risk of re-revision following secondary patellar resurfacing. A retrospective observational study of the National Joint Registry (NJR) was performed. All primary TKRs for osteoarthritis alone performed between 1 April 2003 and 31 December 2016 were eligible (n = 842,072). Patellar resurfacing during TKR was performed in 36% (n = 305,844). The primary outcome was all-cause revision surgery. Secondary outcomes were the number of excess all-cause revisions associated with using TKRs without (versus with) patellar resurfacing, and the risk of re-revision after secondary patellar resurfacing.Aims
Methods
In this phase 2 clinical trial (EudraCT 2011-000541-20) we examined the effect of denosumab versus placebo on osteolytic lesion activity in patients undergoing revision surgery after THA. Men and women ≥ 30 years old scheduled for revision surgery for symptomatic, radiologically-confirmed osteolysis were randomised (1:1) to receive either denosumab 60mg or placebo subcutaneously eight weeks prior to operation. At surgery, biopsies from the osteolytic membrane-bone interface were taken for histomorphometric analysis of osteoclast number, the primary outcome measure. Secondary outcome measures included other static histomorphometric indices and systemic bone turnover markers. Adverse events and patient-reported clinical outcome scores were recorded as safety endpoints.Abstract
Objective
Methods
To develop and validate patient-centred algorithms that estimate individual risk of death over the first year after elective joint arthroplasty surgery for osteoarthritis. A total of 763,213 hip and knee joint arthroplasty episodes recorded in the National Joint Registry for England and Wales (NJR) and 105,407 episodes from the Norwegian Arthroplasty Register were used to model individual mortality risk over the first year after surgery using flexible parametric survival regression.Aims
Methods
The use of cementless acetabular components is currently the gold standard for treatment in total hip arthroplasty (THA). Porous coated cups have a low modulus of elasticity that enhances press-fit and a surface that promotes osseointegration. Monoblock acetabular cups represent a subtype of uncemented cup with the liner moulded into the metal shell, minimizing potential backside wear and eliminating the chance of mal-seating. The aim of this study was to compare the short-term clinical and radiographic performances of a modular cup with that of a monoblock cup, with particular interest in the advent of lucent lines and their correlation with clinical outcomes. In this multi-surgeon, prospective, randomized, controlled trial, 86 patients undergoing unilateral THA were recruited. Participants were randomized to either a porous-coated, modular metal-on-polyethylene (MoP) acetabular component (n=46) or a hydroxyapatite (HA)- and titanium-coated monoblock shell with ceramic-on-ceramic (CoC) bearing (n=42). The porous-coated cup had an average pore size of 250 microns with an average volume porosity of 45%, whereas the monoblock shell had an average pore size of 300 microns with an average volume porosity of 48% and a HA coating thickness of 80 nm. There were no baseline demographic differences between both groups regarding sex, age, body mass index (BMI), or American Society of Anaesthesia (ASA) class (p>0.05). All of the sockets were under-reamed by 1 mm. Radiographs and patient-reported outcome measures (PROMs), including modified Harris Hip Score (mHHS), Western Ontario and McMaster Universities Arthritis Index (WOMAC) and University of California at Los Angeles (UCLA) Hip Score, were available for evaluation at a minimum of 2 years of follow-up. A radiolucent distance between the cup and acetabulum of ≥0.5 mm was defined as gap if it was diagnosed from outset or as radiolucency if it had sclerotic edges and was found on progressive x-ray analyses. Thirty-two gaps (69%) were found in the modular cup group and 28 (6%) in the monoblock one (p=0.001). Of the former, 17 filled the gaps whereas 15 turned into a radiolucency at final assessment. Of the latter, only 1 of the gaps turned into a radiolucency at final follow-up (p 0.05) in both groups. Only the porous-coated cup was an independent predictor of lucent lines (OR:0.052, p=0.007). No case underwent revision surgery due to acetabular loosening during the study period. Only 2 cases of squeaking were reported in the CoC monoblock shell. Both porous-coated modular and hydroxyapatite-coated monoblock cups showed successful clinical results at short-term follow-up, however, the former evidenced a significantly higher rate of radiolucent line occurrence, without any association with PROMs. Since these lines indicate the possibility of future cup loosening, longer follow-up and assessment are necessary.
Local and systemic concentrations of cobalt (Co) and chromium (Cr) ions may be elevated in patients with accelerated tribo-corrosion at prosthesis bearing surfaces and modular taper junctions. Previous studies by us and others have shown that exposure to these metals negatively affect the viability and function of osteoblasts and osteoclasts in vitro, with implications for bone health. More recently, we have observed an increase in total bone mineral density and reduced bone turnover (TRAP5b and osteocalcin) in patients with well-functioning metal-on-metal hip resurfacing (MOMHR). Here, we provide data to support the hypothesis that osteoclast differentiation and function is altered in this patient population, and that this effect is transferrable through their serum. Patients with well-functioning MOMHR (n=18) at median follow-up of 8 years were individually matched for gender, age and time-since-surgery to a low-exposure group consisting of patients with metal-on-polyethylene total hip arthroplasty (THA). The median circulating concentrations of Co and Cr for the MOMHR group were 2.53µg/L and 2.5µg/L respectively, compared to 0.02µg/L and 0.03µg/L for the THA group. Monocyte fraction of peripheral blood was isolated from these patients, seeded onto dentine wafers and differentiated into osteoclasts using media supplemented with RANKL and M-CSF (osteoclastogenic media, OM). Cultures were monitored for the onset of resorption, following which they were treated with OM, autologous serum or serum from the other individual within the matched MOMHR - THA pair, all supplemented with RANKL and M-CSF. At the end of the culture, cells were TRAP stained and quantified for total osteoclast number, number of resorbing osteoclasts and percentage resorption using the CellD Software Package (Olympus, Southend-on-Sea, U.K.). For cells differentiated in osteoclastogenic media, the resorbing ability of osteoclasts derived from MOMHR patients was reduced by 30% (P=0.046) compared to THA. Correlation analyses showed that chronic exposure to Co and Cr trends towards negative association with resorption ability of these osteoclasts (r = −0.3, P=0.06). Furthermore, the resorbing ability of osteoclasts generated from MOMHR patients and differentiated in autologous serum was reduced 33% (p < 0 .0001), whilst matched THA serum caused a smaller reduction of 14% (p < 0 .01). When cells derived from THA patients were differentiated in autologous serum, the resorbing ability of osteoclasts was similarly reduced by 35% (p < 0 .0001), whilst the matched MOMHR serum also caused a reduction of 21% (p < 0 .0001). Reduced osteoclastogenic response of precursor cells from patients with higher circulating Co and Cr suggests an inherent change in their potential to differentiate into functional osteoclasts. The data also suggests that functional response of mature osteoclasts generated from patient precursor cells are dependent on the prior systemic metal concentrations and the presence of higher circulating CoCr in patients with MOMHR. These effects are modest, but may explain the subtle increase in systemic bone mineral density and decreased bone turnover observed in patients after 8 years exposure compared to age, sex, and exposure-time matched patients who received a conventional THA.
In conventional DXA (Dual-energy X-ray Absorptiometry) analysis, pixel bone mineral density (BMD) is often averaged at the femoral neck. Neck BMD constitutes the basis for osteoporosis diagnosis and fracture risk assessment. This data averaging, however, limits our understanding of localised spatial BMD patterns that could potentially enhance fracture prediction. DXA region free analysis (RFA) is a validated toolkit for pixel-level BMD analysis. We have previously deployed this toolkit to develop a spatio-temporal atlas of BMD ageing in the femur. This study aims first to introduce bone age to reflect the overall bone structural evolution with ageing, and second to quantify fracture-specific patterns in the femur. The study dataset comprised 4933 femoral DXA scans from White British women aged 75 years or older. The total number of fractures was 684, of which 178 were reported at the hip within a follow-up period of five years. BMD maps were computed using the RFA toolkit. For each BMD map, bone age was defined as the age for which the L2-norm between the map and the median atlas at that age is minimised. Next, bone maps were normalised for the estimated bone age. A t-test followed by false discovery rate (FDR) analysis was applied to compare between fracture and non-fracture groups. Excluding the ageing effect revealed subtle localised patterns of loss in BMD oriented in the same direction as principal tensile curves. A new score called f-score was defined by averaging the normalised pixel BMD values over the region with FDR q-value less than 1e–6. The area under the curve (AUC) was 0.731 (95% confidence interval (CI)=0.689–0.761) and 0.736 (95% CI=0.694–0.769) for neck BMD and f-score. Combining bone age and f-score improved the AUC significantly by 3% (AUC=0.761, 95% CI=0.756–0.768) over the neck BMD alone (AUC=0.731, 95% CI=0.726–0.737). This technique shows promise in characterizing spatially-complex BMD changes, for which the conventional region-based technique is insensitive. DXA RFA shows promise to further improve fracture prediction using spatial BMD distribution.
Ageing is associated with a gradual and progressive bone loss, which predisposes to osteoporosis. Given the close relationship between the involutional bone loss and the underlying mechanism of osteoporosis, improving the understanding of the bone ageing process can lead to enhanced preventive and therapeutic strategies for osteoporosis. To facilitate this understanding, we develop a spatio-temporal atlas of ageing bone in the femur. We applied our method to a cohort of 11,576 Caucasian women (20–97 years). We amalgamated data from three different studies: 5095 women from the UK Biobank study, 1609 women from the OPUS study, and 5112 women from the MRC-Hip study. The scans are collected using either a Hologic QDR 4500A (Waltham, MA), a Lunar GE iDXA (Madison, WI), or a Lunar GE Prodigy (Madison, WI). Pixel BMD maps were exported using APEX v3.2 and Encore v16 for scans collected on Hologic Inc. and Lunar Corp., respectively. The method utilises a thin plate spline (TPS) registration to warp each scan to a reference mean shape. This image warping, termed Region Free Analysis (RFA), aims to eliminate morphological variation and establish a correspondence between pixel coordinates. At each pixel coordinate, the BMD evolution with ageing was modelled using smooth quantile curves. We deployed the R-package ‘VGAM’ to fit the smooth quantile curves. Cortical thinning was observed consistently with ageing around the shaft from the 60th onwards. A widespread bone loss was also observed in the trochanteric area. Quantile regression curves demonstrated different rates of bone loss at different anatomic locations. For example, bone loss was observed consistently in the mid-femoral neck, while bone mass was preserved the most in the inferior cortex. The developed atlas provides new insights into the spatial bone loss patterns, for which the conventional DXA analysis is insensitive.
The development of an algorithm that provides accurate individualised estimates of revision risk could help patients make informed surgical treatment choices. This requires building a survival model based on fixed and modifiable risk factors that predict outcome at the individual level. Here we compare different survival models for predicting prosthesis survivorship after hip replacement for osteoarthritis using data from the National Joint Registry for England, Wales, Northern Ireland and the Isle of Man. In this comparative study we implemented parametric and flexible parametric (FP) methods and random survival forests (RSF). The overall performance of the parametric models was compared using Akaike information criterion (AIC). The preferred parametric model and the RSF algorithm were further compared in terms of the Brier score, concordance index (C index) and calibration. The dataset contains 327 238 hip replacements for osteoarthritis carried out in England and Wales between 2003 and 2015. The AIC value for the FP model was the lowest. The averages of survival probability estimates were in good agreement with the observed values for the FP model and the RSF algorithm. The integrated Brier score of the FP model and the RSF approach over 10 years were similar: 0.011 (95% confidence interval: 0.011–0.011). The C index of the FP model at 10 years was 59.4% (95% confidence interval: 59.4%–59.4%). This was 56.2% (56.1%–56.3%) for the RSF method. The FP model outperformed other commonly used survival models across chosen validation criteria. However, it does not provide high discriminatory power at the individual level. Models with more comprehensive risk adjustment may provide additional insights for individual risk.
Heterotopic ossification (HO) is lamellar bone formation that occurs within tissues that do not normally have properties of ossification. The pathoaetiology of HO is poorly understood. We conducted a genome wide association study to better understand the genetic architecture of HO. 891 patients of European descent (410 HO cases) following THA for primary osteoarthritis were recruited from the UK. HO was assessed from plain AP radiographs of the pelvis. Genomic DNA was extracted, genotyped using the Illumina 610 beadchip and referenced using the 1000 Genome Project panel. HO susceptibility case-control analysis and an evaluation of disease severity in those with HO was undertaken using SNPTESTv2.3.0 on>10 million variants. We tested variants most strongly associated with HO in an independent UK THA replication cohort comprising 209 cases and 211 controls. The datasets were meta-analysed using PLINK. In the discovery cohort 70 signals with an index variant at p<9×10–5 were suggestively associated with HO susceptibility. The strongest signal lay just downstream of the gene ARHGAP18 (rs59084763, effect allele frequency (EAF) 0.19, OR1.87 [1.48–2.38], p=2.48×10–8), the second strongest signal lay within the long non-coding (LNC) RNA gene CASC20 (rs11699612, EAF 0.25, OR1.73 [1.1.40–2.16, p=9.3×10–8). In the discovery cohort 73 signals with an index variant at p<9×10–5 were associated with HO severity. At replication, 12 of the leading 14 susceptibility signals showed a concordant direction of allelic effect and 5 replicated at nominal significance. Following meta-analysis, the lead replicating susceptibility signal was the CASC20 variant rs11699612 (p=2.71×10–11). We identify consistent replicating association of variation within the LNC RNA CASC20 with HO susceptibility after THA. Although the function of CASC20 is currently unknown, possible mechanisms include transcriptional, post-transcriptional and epigenetic regulation of downstream target genes. The work presented here provides new avenues for the development of novel predictive and therapeutic approaches towards HO.
Commonly used alterations of prosthetic surfaces include grit-blasting (GB), plasma-sprayed titanium (Ti) or hydroxyapatite (HA) coating. Systemic concentrations of cobalt (Co) and chromium (Cr) are elevated in patients with metal-on-metal hip replacement, but can occur for all modular hip replacements. Here, we use whole genome microarrays to assess differential gene expression in primary human osteoblasts grown in vitro and on these prosthesis surfaces following exposure to clinically relevant concentrations of Co and Cr. Mesenchymal cells obtained from bone-fragments of 3 patients undergoing joint replacement surgery were differentiated into osteoblasts. Subsequently, cells were cultured in vitro on tissue-culture plates (TCP), or on GB, Ti and HA surfaces (JRI Orthopaedics Ltd, Sheffield, UK). Following 24hr exposure to a combination of clinically equivalent concentrations of Co2+:Cr3+, RNA was extracted and hybridized to SurePrint-G3 Gene Expression Microarray. Probe signals were normalised using ‘Limma’ package on R-Bioconductor and differential gene expression assessed with empirical Bayes approach (Log2FC>1.00, P<0.001 for differentially expressed genes). For cells grown on TCP, 11 genes were upregulated with 500μg/L Co2+:Cr3+. Of these, 4 were associated to HIF-1 signalling based on KEGG pathway analysis (P=5.4e-5). Exposure to 1000μg/L Co2+:Cr3+ altered expression at 164 loci for HA surfaces, and a separate 50 loci for Ti surfaces compared to GB surfaces. Genes for osteoblast differentiation (BMP2 and RGS2) were downregulated on HA surfaces compared to GB, whilst genes for cell-adhesion (ESAM), vesicular trafficking (RAB37) and protection against oxidative damage (NRF2) were upregulated. Ti surfaces caused an upregulation in ERBB3 and CNTF, which are associated with inhibition of osteoblast differentiation and mineralisation, when compared to GB surfaces. This study confirms the role of HIF-1 signalling in response to prosthesis generated metal ions, and is the first to provide a comprehensive genome-wide insight into transcriptional response of osteoblasts at prosthesis surface to clinically equivalent metal exposure.
The aim of this study was to describe temporal trends and survivorship of total hip arthroplasty (THA) in very young patients, aged ≤ 20 years. A descriptive observational study was undertaken using data from the National Joint Registry (NJR) for England, Wales, Northern Ireland and the Isle of Man between April 2003 and March 2017. All patients aged ≤ 20 years at the time of THA were included and the primary outcome was revision surgery. Descriptive statistics were used to summarize the data and Kaplan–Meier estimates calculated for the cumulative implant survival.Aims
Patients and Methods
The aim of this study was to determine whether patients with
metal-on-metal (MoM) arthroplasties of the hip have an increased
risk of cardiac failure compared with those with alternative types
of arthroplasties (non-MoM). A linkage study between the National Joint Registry, Hospital
Episodes Statistics and records of the Office for National Statistics
on deaths was undertaken. Patients who underwent elective total
hip arthroplasty between January 2003 and December 2014 with no
past history of cardiac failure were included and stratified as
having either a MoM (n = 53 529) or a non-MoM (n = 482 247) arthroplasty.
The primary outcome measure was the time to an admission to hospital
for cardiac failure or death. Analysis was carried out using data
from all patients and from those matched by propensity score.Aims
Patients and Methods