16 to 34% of the population suffer from shoulder pain, the most common cause being rotator cuff tears. NICE guidance recommends using ultrasound scan (USS) or MRI to assess these patients, but does not specify which is preferable. This study assesses the accuracy of USS and MRI in rotator cuff tears in a DGH, to establish the most appropriate imaging modality. Patients who had at least two of shoulder ultrasound, MRI or arthroscopy within a seven month period (n=55) were included in this retrospective study. Sensitivity, Specificity, Positive Predictive Value (PPV) and Negative Predictive Value (NPV) were calculated using arthroscopy as the true result, and kappa coefficients calculated for each pairing. 59 comparisons were made in total. Sensitivity for MRI in full supraspinatus tears was 0.83, and for USS 0.75. Specificity for MRI in these tears was 0.75, and for USS 0.83. Values were much lower in other tears, which occurred less frequently. USS and MRI completely agreed with each other 61.3% of the time. Both modalities were only completely accurate 50% of the time. Kappa coefficient between arthroscopy and MRI for supraspinatus tears was 0.658, and for USS was 0.615. There was no statistical difference between MRI and USS sensitivity or specificity (p=1), suggesting that one modality cannot be recommended over the other for full supraspinatus tears. They also do not tend to corroborate one another, suggesting that there is no benefit from doing both scans. Further research is needed to see how both modalities can be improved to increase their accuracy

FAI has been implicated in the progression of osteoarthritis (OA) and early detection may allow for treatment, which can slow or halt progression. FAI is a difficult condition to image and there is little objective evidence about imaging accuracy. We aim to measure the accuracy of five imaging modalities. Three blinded observers retrospectively reviewed five different modalities from two age and sex matched groups: A patient group referred to the outpatient clinic with a clinical diagnosis of FAI and a control group who had had CT scans of the pelvis for suspected trauma, where the Pelvic scan had been reported as showing no injuries. The imaging modalities were: Standard x-ray; Antero-Posterior, Lateral; Condition-specific x-ray projections; Dunn view, lateral internal rotation; Standard Computer Tomography (CT) multiplanar reconstruction (MPR); axial, sagittal and coronal; Condition-specific CT MPR; angled axial, angled coronal; 3D modelling; and surface rendered dynamic. We found marked variations in the sensitivity, specificity, Positive Predictive Value (PPV) and Negative Predictiive Value (NPV) for each of the following imaging modalities: Standard X-ray; Sensitivity 51.9; Specificity; 57.1; PPV; 40; NPV; 68.3 Special X-rays; Sensitivity; 66.7; Specificity; 57.1; PPV; 46.1; NPV; 75.7. Standard CT MPR; Sensitivity; 40.7; Specificity; 75.5; PPV; 47.8; NPV; 69.8 Special CT MPR; Sensitivity; 48.1; Specificity; 57.1; PPV; 46.4; NPV; 70.8 Dynamic 3D CT models; Sensitivity; 55.6; Specificity; 69.3; PPV; 42.8; and NPV; 71.8. The Dynamic 3D CT models (where the observer can manipulate the model in real time three dimension to control the perspective) proved to be the most accurate, closely followed by the special X-Ray views, which were also the most sensitive. The Standard CT MPRs were the most specific but had a low sensitivity. This is the first study to measure sensitivity, specificity and PPV and NPV for these imaging modalities in FAI. We recommend the use of condition-specific X-Ray views as well as 3D CT Models for optimal imaging accuracy in this condition. Standard X-Ray views and CTs proved less useful

Magnetic resonance imaging (MRI) is a useful diagnostic tool in evaluating meniscus pathology in the knee. Data from available literature suggests sensitivity and specificity rates around 90% when compared to the gold standard findings at knee arthroscopy. We sought to evaluate the sensitivity, specificity and precision rate (positive predictive value) of MRI at diagnosing meniscus tears within our unit. A retrospective audit of a total of 79 MRI reports and arthroscopic findings spanning a one year period was carried out. There were 66 positive MRI reports and 13 negative reports. There were 6 false positives 4 false negatives when compared to arthroscopic findings. The sensitivity of MRI for detecting meniscus tears was 93.7% with 60 out of 64 tears detected. All 4 false negatives also had at least grade III osteoarthritic changes at arthroscopy. Specificity was rather low at 60% with MRI reporting 6 tears (false positives) out of 15 patients who had no tears found at arthroscopy. The positive predictive value (precision rate) of MRI detecting tears was 90.9%. This data shows that MRI in our unit has a comparable high sensitivity to that in various literature making it a useful tool at ruling out disease with a negative result in the clinical setting. A more useful parameter in the clinical setting is its high precision rate when faced with a positive result. However, its specificity is much lower than that in most published data. A total of 6 tears on MRI turned out not to be on arthroscopy meaning patients could have been subjected to an avoidable invasive procedure in the absence of any other indication. This highlights the importance of obtaining reports from experienced musculoskeletal radiologists and the need for surgeons to review MRI images and match them to clinical information prior to subjecting patients to surgery

Aims: Multiple nerve repair by means of a Y-shaped nerve guide represents a good model for studying the specificity of peripheral nerve fiber regeneration. Here we have employed this model for investigating the specificity of axonal regeneration in mixed nerves of the rat forelimb model. Specificity of nerve regeneration can be defined as the ability of the nerve fibers of a peripheral nerve, after a lesion. Tree types of specificity on nerve regeneration has been postulated: “tissue specificity” (the preferential reinnervation of distal nerve tissue versus other types of tissue), topographic specificity (regenerating nerve fibers are preferentially attracted by analogous distal pathways (e.g. preferential regeneration along tibial nerve pathways by tibial nerve fibers), and end-organ specificity, which is the hypothesis that distal end-organs (muscle vs. sensory targets) specifically attracts the respective (motor vs. sensory) regenerating nerve fibers. Exists no agreement regarding the presence and features of the two last type of specificity. Methods: The left median and ulnar nerves, in adult female rats, were transected and repaired with a 14-mm Y-shaped conduit. The proximal end of the Y-shaped conduit was sutured to the proximal stump of either the median nerve or the ulnar nerve. Ten months after surgery, rats were tested for functional recovery of each median and ulnar nerve. Quantitative morphology of regenerated myelinated nerve fibers was then carried out by the two-dimensional disector technique. Results: Results showed that partial recovery of both median and ulnar nerve motor function was regained in all experimental groups. Performance in the grasping test was significantly lower when the ulnar nerve was used as the proximal stump. Ulnar test assessment showed no significant difference between the two Y-shaped repair groups. The number of regenerated nerve fibers was significantly higher in the median nerve irrespectively of the donor nerve, maintaining the same proportion of myelinated fibers between the two nerves (about 60% median and 40% ulnar). On the other hand, nerve fiber size and myelin thickness were significantly larger in both distal nerves when the median nerve was used as the proximal donor nerve stump. G-ratio and myelin thickness/ axon diameter ratio returned to normal values in all experimental groups. Conlusions: These results demonstrate that combined Y-shaped-tubulization repair of median and ulnar nerves permits the functional recovery of both nerves, independently from the proximal donor nerve employed, and that tissue, and not topographic, specificity guides nerve fiber regeneration in major forelimb mixed nerves of rats

Abstract

There is a need for the accurate measurement of the outcome after knee surgery. The Oxford Knee Score is being increasingly used since it is reported to be short, simple, inexpensive and validated.

We sent the questionnaire to 346 patients awaiting surgery to the hip or lumbar spine. Only 11% of 141 patients with proximal pathology who denied knee problems gave a maximum score. Their mean score was substantially lower than expected at 28.7 (maximum 48), and was significantly lower than the score of 36.5 obtained from patients after total knee replacement.

We therefore suggest that the frequent coexistence of hip or spinal pathology will significantly alter both the absolute score and any improvement to be expected after knee surgery. Although sensitive to disability originating from the knee the Oxford Knee Score is not sufficiently specific since it is heavily influenced by more proximal pathology.

Aims

The sensitivity and specificity of electrodiagnostic parameters in diagnosing carpal tunnel syndrome (CTS) have been reported differently, and this study aims to address this gap.

Aims

This study aimed to answer two questions: what are the best diagnostic methods for diagnosing bacterial arthritis of a native joint?; and what are the most commonly used definitions for bacterial arthritis of a native joint?

Aims: We aim to investigate how often we needed extensive investigations to reach a speciþc diagnosis and whether we could identify symptoms and signs predictive of such diagnosis. Methods: We analysed the records of 60 children referred to our specialist clinics complaining of back pain. We statistically analysed the speciþcity of 9 parameters from history and clinical examination to predict the þnal diagnosis and checked the correlation between initial clinical diagnosis and þnal diagnosis following extensive investigations such as bone scan, CT and MRI scan. Results: The mean age of our group of patients was 12.3 years (range=3–18 years). We failed to reach a speciþc initial diagnosis after history taking, clinical examination and simple X-rays of the spine in 33 patients (55%). By the completion of investigations a speciþc diagnosis was not reached in only 18 (35%), and the difference was signiþcant compared to the other 65% (chi square(1)=4.41, p< 0.05). In 15 children specialist investigations clariþed a diagnosis of spondylolysis, Scheuermannñs disease and tumour. A hypothesised set of 9 clinical symptoms and signs was not predictive of speciþcity of initial clinical diagnosis (RL 2=0.12, chi square (9)=10.07, p> 0.05), but was predictive of þnal diagnosis (RL 2=0.35, chi square(9)=22.88, p< 0.01). A speciþc diagnosis was less likely in the presence of activity related pain or intermittent pain. There was a high degree of correlation between initial clinical diagnosis and þnal diagnosis. (phi=0.70, chi square(1)=25.07, p< 0.001). Conclusions: For about half our group of children, back pain was not a speciþc symptom. A set of common set of clinical symptoms and signs did not initially predict speciþcity. Based on these initial þndings we should continue to investigate in depth any referred child with back pain.

Background: The diagnosis of implant-associated infections is difficult due to organisms attached to surfaces as biofilms. We hypothesize that diagnosis can be improved by removing biofilm microorganisms from implant surface by sonication, followed by Gram stain, culture and calorimetric detection in sonication fluid.

Methods: We prospectively included adult patients from May 2005 until December 2006 from whom an orthopedic implant (joint prosthesis or internal fixation device) was removed for any reason. Removed implants were vortexed and sonicated in solid containers 5 min at 40 kHz in 100 to 400 ml Ringer’s solution. The resulting sonicate was plated and incubated on aerobic and anaerobic blood agar and aliquots were in parallel incubated at 37°C for 3 days in an isothermal calorimeter TAM III (TA Instruments, New Castle, DE). Gram stain was performed on sonicates centrifuged at 5000 g for 10 min. Definitive infection diagnosis was of the implant was defined if purulence surrounding the implant, or growth of the same microorganism in ≥2 synovial fluid or intraoperative tissue specimens, or acute inflammation in histopathology, or a sinus tract was present. Sonicate culture was defined positive if > 10 cfu (colony forming units) grew/ml sonicate. Calorimetry was defined positive if heat flow rate increased ≥10 μW above baseline (detection limit ~0.3 μW).

Results: 846 implants (367 joint prostheses and 479 internal fixation devices) were studied, of which 171 (20%) were infected and 675 (80%) were aseptic cases. The sensitivity of intraoperative tissue cultures was 74%, of sonicate culture 89%, of sonicate Gram stain 51%, of sonicate calorimetry 96%. The specificity of all specimens was ≥95%.

Conclusion: Sonicate culture and calorimetry were more sensitive than intraoperative tissue cultures for diagnosing implant infections. With Gram stain of centrifuged sonicate, infection was diagnosed in > 50% cases. Sonicate culture and calorimetry may replace the current approach using multiple intraoperative periprosthetic tissue specimens, whenever the implant or part of it is removed.

ARMD (Adverse Reaction to Metal Debris) is an increasingly recognised complication of metal on metal hip replacements. The MHRA (Medical and Healthcare Related Devices Agency) have advised a blood cobalt or chromium level above 7 mg/L is a threshold for further investigation, stating that “low levels are reassuring and strongly predict not having an adverse outcome”. Cross-sectional imaging should be performed when levels are above 7 mg/L. We have performed a study investigating the specificity and sensitivity of chromium and cobalt metal ion levels as a screening measure for ARMD.

79 ASR hip replacements were performed at our hospital and 75 (95%) of these underwent a Metal Artefact Reduction Sequence (MARS) MRI scan. All patients (64 hips) who had not undergone revision were invited to take part in this study. 57 patients with 62 hip replacements completed hip and activity scores, had blood cobalt and chromium ion level measurements and 3D-CT to measure acetabular component position.

Acetabular component inclination (>50 degrees), small head size (< 51mm) and female gender were significantly correlated with raised chromium (Cr) and cobalt (Co) ion levels. An ARMD was detected using MRI in 18 (29%) of the hips in this study. The incidence of ARMD was significantly higher when chromium concentration was above 7 mg/l (p = 0.02). Chromium ion levels >7 mg/L had a sensitivity of 56% and specificity of 83% for ARMD, and cobalt ion levels >7 mg/L 56% and 76% respectively. 40 patients had cobalt levels <7 mg/L and 33 had chromium levels <7 mg/L, but 8 of these had an ARMD on MRI. All 8 patients had minimal symptoms (Oxford Hip Score ≥ 44 out of 48).

The Medicines and Healthcare Products Regulatory Agency (MHRA) has recommended that cobalt and chromium levels be measured in patients with a metal-on-metal hip replacement and cross-sectional imaging performed when these levels are above 7 μg/L. This study has shown that by using this threshold, in patients with this implant combination, the sensitivity and specificity for the detection of ARMD is low and patients with soft tissue disease may be missed. Furthermore the presence of MRI detected ARMD, in the absence of significant clinical symptoms and with metal ion levels <7 μg/L is of concern.

MoM implants at risk of failure are associated with raised cobalt and chromium levels. However metal ion analysis alone is not reliable as a screening tool for ARMD, which is often clinically “silent”. We recommend the routine use of MARS MRI as the safest method of ARMD diagnosis in patients with MoM implants.

Aims

The risk to patients and healthcare workers of resuming elective orthopaedic surgery following the peak of the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) pandemic has been difficult to quantify. This has prompted governing bodies to adopt a cautious approach that may be impractical and financially unsustainable. The lack of evidence has made it impossible for surgeons to give patients an informed perspective of the consequences of elective surgery in the presence of SARS-CoV-2. This study aims to determine, for the UK population, the probability of a patient being admitted with an undetected SARS-CoV-2 infection and their resulting risk of death; taking into consideration the current disease prevalence, reverse transcription-polymerase chain reaction (RT-PCR) testing, and preassessment pathway.

Plasma levels of cobalt and chromium ions and Metal Artefact Reduction Sequence (MARS)-MRI scans were performed on patients with 209 consecutive, unilateral, symptomatic metal-on-metal (MoM) hip arthroplasties. There was wide variation in plasma cobalt and chromium levels, and MARS-MRI scans were positive for adverse reaction to metal debris (ARMD) in 84 hips (40%). There was a significant difference in the median plasma cobalt and chromium levels between those with positive and negative MARS-MRI scans (p < 0.001). Compared with MARS-MRI as the potential reference standard for the diagnosis of ARMD, the sensitivity of metal ion analysis for cobalt or chromium with a cut-off of > 7 µg/l was 57%. The specificity was 65%, positive predictive value was 52% and the negative predictive value was 69% in symptomatic patients. A lowered threshold of > 3.5 µg/l for cobalt and chromium ion levels improved the sensitivity and negative predictive value to 86% and 74% but at the expense of specificity (27%) and positive predictive value (44%).

Metal ion analysis is not recommended as a sole indirect screening test in the surveillance of symptomatic patients with a MoM arthroplasty. The investigating clinicians should have a low threshold for obtaining cross-sectional imaging in these patients, even in the presence of low plasma metal ion levels.

Background. Masses are not uncommon in the foot and ankle. Most of these masses are benign, often leading clinicians to underestimate their potential for malignancy. Methods. We conducted a retrospective review of our clinical records, on patients with histologically confirmed musculoskeletal tumours of the foot and ankle, treated in a single nstitution between 2010 and 2019. The maximum diameter of each lesion was determined through MRI or Ultrasound analysis in centimeters. To develop a scoring system we compared the risk of malignancy with five criteria: site (proximal or distal to the first TMTJ), gender, age, composition and the diameter as observed. Results. Our study included 496 patients, of whom 39 (7.9%) were identified as having malignancies. The incidence of malignancy demonstrated an increased propensity among male patients, patients over 50 years of age and lesions located proximal to the TMTJ. A ROC Analysis determined that lesions measuring over 2.85 cm had an increased risk of malignancy, with a PPV of 31.1%, a NPV of 94.2%, a Sensitivity of 0.82, and a Specificity of 0.62. These identified patterns of risk were employed to formulate a scoring system, aimed at facilitating informed clinical judgment in the referral of patients to regional tumor services. Conclusion. The new OxFAT scoring system highlights the importance of lesion size, site, age and gender of the patient in determining the risk of malignancy in lump in the foot and ankle. We propose this new scoring system to aid health care professionals in managing these patients. Based on our results any patient with a foot or ankle mass of less than 2.85cm, an OxFAT score < 4/7 and no malignant or sinister features on MRI or USS can be managed locally with excision biopsy. All other patients should be referred urgently to a Regional Tumour Service

Introduction and purpose: To assess the diagnostic usefulness of MRI to diagnose ACL lesions using quantitative instead of qualitative parameters. Materials and methods: A retrospective study was made of the MRIs of a group of 50 patients with an athroscopically confirmed ACL tear and a control group of 50 patients with meniscal lesions and with an arthroscopic confirmation that they had a normal ACL. Multiple MRI findings were studied in order to evaluate their sensitivity and Specificity to detect an ACL lesion. Special emphasis was placed on 3 quantitative parameters, including a simplified method to measure the angle between the ACL and the tibial plateau. Results: Using 45° as a cutoff value, the ACL/tibial plateau angle yielded a sensitivity and Specificity of 100%. With a 0° cutoff value, the angle together with Blumensaat’s line showed a sensitivity of 90% and a Specificity of 98%. With a cutoff value of 115°, the PCL angle showed a sensitivity of 70% and a Specificity of 82%. Conclusions: The quantitative parameters studied are valuable to predict ACL lesions and can increase both the sensitivity and Specificity of MRI images. The ACL/ tibial plateau angle can be measured easily using a single MRI image and can be regarded as the best clue to diagnose ACL lesions

Introduction: Ankle sprain is one of the common presentations in Accident and Emergency. Accurate diagnosis is critical and sometime difficult without special investigation such as Magnetic Resonance Imaging (MRI) or arthroscopy. Aim: The purpose of this project was to evaluate magnetic resonance imaging (MRI) accuracy for diagnosing and defining ankle pathology. Method: Retrospective review of 36 patients who underwent both MRI and Arthroscopy. All cases were seen by single orthopaedic surgeon with special interest in foot and ankle surgery. MRI scan were reviewed by consultant radiologist at our institute. Arthroscopy was used as a standard for comparing MRI results. Sensitivity and Specificity was calculated by qualified statistician. Results: For osteochondral lesion the Sensitivity of 85.7 %, Specificity of 93.3% and Accuracy of 89.7% was noticed. Anterior and posterior tibiofibular ligament (ATFL and PTFL) pathology had a Sensitivity of 100 %, Specificity of 100% and Accuracy of 100%, while anterior inferior tibiofibular ligament pathology had a Sensitivity of 66.6 %, Specificity of 95% and Accuracy of 86.6%. Conclusion: Although MRI is a useful tool in exclusion of pathological condition its sensitivity and accuracy in diagnosing ligament injury is not encouraging

Aim. The clinical relevance of microbial DNA detected via next-generation sequencing (NGS) remains unknown. This multicenter study was conceived to: 1) identify species on NGS that may predict periprosthetic joint infection (PJI), then 2) build a predictive model for PJI in a developmental cohort, and 3) validate predictive utility of the model in a separate multi-institutional cohort. Method. Fifteen institutions prospectively collected samples from 194 revision TKA and 184 revision THA between 2017–2019. Synovial fluid, tissue and swabs were obtained intraoperatively and sent to MicrogenDx (Lubbock, TX) for NGS analysis. Reimplantations were excluded. Patients were classified per the 2018 ICM definition of PJI. DNA analysis of community similarities (ANCOM) was used to identify 17 bacterial species of 294 (W-value>50) for differentiating infected vs. noninfected cases. Logistic regression with LASSO selection and random-forest algorithms were then used to build a model for predicting PJI. ICM classification was the response variable (gold-standard) and species identified through ANCOM were predictors. Patients were randomly allocated 1:1 into training and validation sets. Using the training set, a model for PJI diagnosis was generated. The entire model-building procedure and validation was iterated 1000 times. Results. The model's assignment accuracy was 75.9%. There was high accuracy in true-negative and false-negative classification using this model, which has previously been a criticism of NGS. Specificity was 97.1%, PPV 75.0% and NPV 76.2%. On comparison of abundance between ICM-positive and ICM-negative patients, Staphylococcus aureus was the strongest contributor (F=0.99) to model predictive power. In contrast, Cutibacterium acnes was less predictive (F=0.309) and abundant across infected and noninfected revisions. Discussion. This is the first study to utilize predictive algorithms on a large multicenter dataset to transform analytic NGS data into a clinically relevant diagnostic model. Our collaborative findings suggest NGS may be an independent adjunct for PJI diagnosis, while also facilitating pathogen identification. Future work applying machine-learning will improve accuracy and utility of NGS

Aim. Synovial fluid investigation is the best alternative to diagnose prosthetic joint infection (PJI) before adequate microbiological/histology sampling during revision surgery. Although accurate preoperative diagnosis is certainly recommended, puncturing every patient before revision arthroplasty raises concerns about safety and feasibility issues especially in difficult to access joint (e.g., hip), that often require OR time and fluoroscopy/ultrasound guidance. Currently there is no clear guidelines regarding optimal indications to perform preoperative joint aspiration to diagnose PJI before revision surgery. The main goal of this study is to determine the accuracy of our institutional criteria using the new European Bone and Joint Infection Society (EBJIS) PJI definition. Method. We retrospectively evaluated every single- or first-stage for presumed aseptic or known infected revision total hip/knee arthroplasty procedures between 2013–2020. Preoperative clinical and laboratory features were systematically scrutinized. Cases with insufficient information for accurate final PJI diagnosis (i.e., no perioperative synovial fluid examination or no multiple cultures including sonication of removed implant) were excluded. Preoperative joint aspiration is recommended in our institution if any of the following criteria are met: 1) elevated CRP and/or ESR; 2) early failure (<2 years) or repeat failure; 3) high clinical suspicion/risk factors are present. Performance of such criteria were compared against final postoperative EBJIS definition PJI diagnosis. Results. A total of 364 revision THAs or TKAs were performed during the study period. After excluding 258 cases with insufficient information, a total of 106 patients were ultimately included. 38 (35,8 %) were classified as confirmed infections, 10 (9.4 %) as likely infected and 58 (54.7%) as infection unlikely. Of those, 37 confirmed infection cases, 9 likely infected cases and 32 infection unlikely cases did have indication for preoperative synovial fluid collection before revision surgery. Institutional criteria showed 95.8 % Sensitivity, 44.83 % Specificity, 92.9 % Negative Predictive Value (NPV) and 59 % Positive Predictive Value (PPV). Conclusions. Sensitivity and NPV of the aforementioned institutional criteria are very high even with the use of the more sensitive EBJIS PJI definition. As such they seem to be a valid alternative in selecting patients that should be punctured before revision arthroplasty. They identify the vast majority of infected patients while saving a significant number of patients from unnecessary procedures

Aim. Metagenomic nanopore sequencing is demonstrating potential as a tool for diagnosis of infections directly from clinical samples. We have previously shown nanopore sequencing can be used to determine the causative bacterial species in prosthetic joint infections (PJI). However, to make predictions regarding antimicrobial resistance, human DNA contamination must be reduced so a greater proportion of sequence data corresponds to the microbial portion of the DNA extract. Here, we utilise selective DNA extraction from sonication fluid samples to begin to make predictions regarding antimicrobial resistance in PJI. Method. We investigated host cell DNA depletion with 5% saponin selective human cell lysis followed by nuclease digestion. Subsequently, bacterial cells were mechanically lysed before DNA extraction. Sequencing libraries from samples treated with and without saponin were prepared with a Rapid PCR Barcoding Kit. 1. and sequenced in multiplexes of 2–8 samples/flowcell on a GridION. Sequencing reads were analysed using the CRuMPIT pipeline and thresholds to indicate presence of a specific bacterial genus/species were investigated. Antimicrobial resistance determinants were detected using previously published sequences specifically for Staphylococcus aureus, as an example organism frequently causing PJI. Results. 247 DNA extracts from 115 sonication fluids plus controls were subjected to metagenomic sequencing, comprising extracts from 67 culture-positive (10 of which were culture-positive at <50 CFU/ml) and 48 culture-negative samples. 5% saponin depleted human DNA contamination, reducing the number of human sequenced bases to a median 12% from 98% in comparison to 5μm filtration without saponin. In 11 samples 5% saponin depleted human bases by <12% in comparison to 5μm filtration, which may be indicative of incomplete depletion. Bacteria observed in sonication fluid culture were identified to species-level in 49/65 (75%) cases, and to genus-level in 51/65 (78%). Specificity of sequencing was 103/114 (90%). Sequencing made a completely successful prediction of antimicrobial susceptibility in 8/19 S. aureus culture-positive samples treated with 5% saponin, and a partial prediction in 5/19 for the 8 antibiotics investigated. Without 5% saponin treatment sequencing could only detect a limited number of AMR determinants in 3/19 samples. Sequencing correctly predicted 13/15 (87%) resistant and 74/74 (100%) susceptible phenotypes where sufficient sequence data were available. Conclusions. Nanopore metagenomic sequencing can provide species identification in PJI. Additionally, depletion of human DNA improves depth of coverage and allows detection of antimicrobial resistance determinants, demonstrating as a proof of principle that nanopore sequencing could potentially provide a complete diagnostic tool in PJI. 1. Oxford Nanopore Technologies