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

Aims. To evaluate the histopathological examination of peri-implant tissue samples as a technique in the diagnosis of postoperative spinal implant infection (PSII). Methods. This was a retrospective analysis. Patients who underwent revision spinal surgery at our institution were recruited for this study. PSII was diagnosed by clinical signs, histopathology, and microbiological examination of intraoperatively collected samples. Histopathology was defined as the gold standard. The sensitivity for histopathology was calculated. A total of 47 patients with PSII and at least one microbiological and histopathological sample were included in the study. Results. PSII occurred in approximately 28% of the study population. Histopathology showed a sensitivity of 51.1% in the diagnosis of PSII. The most commonly found pathogens were Cutibacterium acnes and gram-positive staphylococci. Conclusion. Histopathology has low sensitivity for detecting PSII. In particular, infections caused by low-virulence microorganisms are insufficiently detected by histopathology. Cite this article: Bone Joint J 2020;102-B(7):899–903

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?. Methods. We performed a search of PubMed, Embase, and Cochrane libraries for relevant studies published between January 1980 and April 2020. Of 3,209 identified studies, we included 27 after full screening. Sensitivity, specificity, area under the curve, and Youden index of diagnostic tests were extracted from included studies. We grouped test characteristics per diagnostic modality. We extracted the definitions used to establish a definitive diagnosis of bacterial arthritis of a native joint per study. Results. Overall, 28 unique diagnostic tests for diagnosing bacterial arthritis of a native joint were identified. The following five tests were deemed most useful: serum ESR (sensitivity: 34% to 100%, specificity: 23% to 93%), serum CRP (sensitivity: 58% to 100%, specificity: 0% to 96%), serum procalcitonin (sensitivity: 0% to 100%, specificity: 68% to 100%), the proportion of synovial polymorphonuclear cells (sensitivity: 42% to 100%, specificity: 54% to 94%), and the gram stain of synovial fluid (sensitivity: 27% to 81%, specificity: 99% to 100%). Conclusion. Diagnostic methods with relatively high sensitivities, such as serum CRP, ESR, and synovial polymorphonuclear cells, are useful for screening. Diagnostic methods with a relatively high specificity, such as serum procalcitonin and synovial fluid gram stain, are useful for establishing a diagnosis of bacterial arthritis. This review helps to interpret the value of various diagnostic tests for diagnosing bacterial arthritis of a native joint in clinical practice. Cite this article: Bone Joint J 2021;103-B(12):1745–1753

Blood metal ions have been widely used to investigate metal-on-metal hip replacements, but their ability to discriminate between well-functioning and failed hips is not known. The Medicines and Healthcare products Regulatory Agency (MHRA) has suggested a cut-off level of 7 parts per billion (ppb). We performed a pair-matched, case-control study to investigate the sensitivity and specificity of blood metal ion levels for diagnosing failure in 176 patients with a unilateral metal-on-metal hip replacement. We recruited 88 cases with a pre-revision, unexplained failed hip and an equal number of matching controls with a well-functioning hip. We investigated the 7 ppb cut-off level for the maximum of cobalt or chromium and determined optimal mathematical cut-off levels from receiver-operating characteristic curves. The 7 ppb cut-off level for the maximum of cobalt or chromium had a specificity of 89% and sensitivity 52% for detecting a pre-operative unexplained failed metal on metal hip replacement. The optimal cut-off level for the maximum of cobalt or chromium was 4.97 ppb and had sensitivity 63% and specificity 86%. Blood metal ions had good discriminant ability to separate failed from well-functioning hip replacements. The MHRA cut-off level of 7 ppb provides a specific test but has poor sensitivity.

A prospective study of allergic contact dermatitis after metal-on-plastic total hip replacement was undertaken in 69 patients, of whom 54 were available for review after operation. Before operation six patients were metal sensitive, but only one remained so afterwards; this patient had given a clear history of metal sensitivity and a titanium prosthesis had therefore been used. In all six patients the result of the operation was good and no case of loosening occurred. Sixty-three patients had negative patch tests before operation; in none of these was metal sensitivity detected after operation. Cutaneous sensitivity to various metals is well documented after the insertion of metal-on-metal prostheses and in failed prostheses. We have not found any such increased sensitivity after metal-on-plastic hip replacement. There is little evidence of a direct causal relationship between metal sensitivity and subsequent loosening. The cutaneous sensitivity may be the consequence of loosening rather than its cause. Our results suggest that, providing metal-on-plastic prostheses are used, routine patch testing before hip replacement is no longer required

Aims. The aims of this study were to investigate any possible relationship between a preoperative sensitivity to pain and the degree of pain at rest and on exertion with postoperative function in patients who underwent stemless total shoulder arthroplasty (TSA). Patients and Methods. In this prospective study, we included 63 patients who underwent stemless TSA and were available for evaluation one year postoperatively. There were 31 women and 32 men; their mean age was 71 years (53 to 89). The pain threshold, which was measured using a Pain Matcher (PM) unit, the degree of pain (visual analogue scale at rest and on exertion, and function using the short version of the Disabilities of the Arm, Shoulder and Hand questionnaire (QuickDASH), were recorded preoperatively, as well as three and 12 months postoperatively. Results. We found an inverse relationship between both the preoperative PM threshold and pain (VAS) at rest and the 12-month postoperative QuickDASH score (Pearson correlation coefficient (r) ≥ 0.4, p < 0.05). A linear regression analysis showed that the preoperative PM threshold on the affected side and preoperative pain (VAS) at rest were the only factors associated with the QuickDASH score at 12 months. Conclusion. These findings indicate the importance of central sensitization in the restoration of function after TSA. Further studies are required to investigate whether extra analgesia and rehabilitation could influence the outcome in at risk patients. Cite this article: Bone Joint J 2018;100-B:480–4

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. Methods. The probability of SARS-CoV-2 infection with a false negative test was calculated using a lower-end RT-PCR sensitivity of 71%, specificity of 95%, and the UK disease prevalence of 0.24% reported in May 2020. Subsequently, a case fatality rate of 20.5% was applied as a worst-case scenario. Results. The probability of SARS-CoV-2 infection with a false negative preoperative test was 0.07% (around 1 in 1,400). The risk of a patient with an undetected infection being admitted for surgery and subsequently dying from the coronavirus disease 2019 (COVID-19) is estimated at approximately 1 in 7,000. However, if an estimate of the current global infection fatality rate (1.04%) is applied, the risk of death would be around 1 in 140,000, at most. This calculation does not take into account the risk of nosocomial infection. Conversely, it does not factor in that patients will also be clinically assessed and asked to self-isolate prior to surgery. Conclusion. Our estimation suggests that the risk of patients being inadvertently admitted with an undetected SARS-CoV-2 infection for elective orthopaedic surgery is relatively low. Accordingly, the risk of death following elective orthopaedic surgery is low, even when applying the worst-case fatality rate. Cite this article: Bone Joint J 2020;102-B(9):1256–1260

1. Evidence is presented which suggests that after total joint replacement bone necrosis and consequent loosening of the prosthesis may be due to the development of sensitivity to the metals used. 2. Nine patients, from a total of fourteen with loose prostheses, were found to be metal sensitive by skin-patch testing. In twenty-four patients with intact prostheses no sensitivity was demonstrated. 3. In material from the joints of sensitive patients the metal content was raised. 4. Examination of this material showed necrosis of bone and soft tissue following obliterative changes in the vascular supply. 5. Similar reactions were found following the injection of cobalt into a sensitive patient. 6. The release of metal around a prosthesis is greatest where metal rubs against itself. 7. We conclude that prostheses in which metal articulates with polyethylene should be preferred; that any patient in whom loosening or fragmentation occurs should be patch tested; and that if sensitivity is found the implant should be removed

Tissues from five patients who underwent revision operations for failed total hip replacements were found to contain large quantities of particulate titanium. In four cases this metal must have come from titanium alloy screws used to fix the acetabular component; in the fifth case it may also have originated from a titanium alloy femoral head. Monoclonal antibody labelling showed abundant macrophages and T-lymphocytes, in the absence of B-lymphocytes, suggesting sensitisation to titanium. Skin patch testing with dilute solutions of titanium salts gave negative results in all five patients. However, two of them had a positive skin test to a titanium-containing ointment.

1. Pain and pressure sensibility has been studied in the fibrous articular ligament and synovial membrane of the knee joint in normal human subjects and in patients subjected to arthrotomy under local anaesthesia.

2. The fibrous ligament was found to be a highly sensitive structure, containing many spots which give rise to sensations of pain or pressure when stimulated mechanically or chemically. Synovial membrane was found to be a relatively insensitive structure which only occasionally contains pain-sensitive spots.

3. Histological examination of articular capsule from the knee joints of normal and sympathectomised cats shows that articular ligaments have a rich nerve plexus and a variety of specialised and unspecialised nerve endings, most of which are somatic in origin. Synovial membrane contains a more delicate nerve network and also a variety of nerve endings, the majority of which are autonomic in origin. But a substantial number of somatic nerves enter the synovial membrane, some of which terminate in nerve loops, globular endings or simple unspecialised endings.

4. Histological examination of normal human articular capsule shows that its innervation closely resembles that of the cat, but no sympathectomised human material was examined.

5. The bearing of these findings on the symptomatology of joint disease is discussed briefly.

Aims. This study aimed to evaluate rasterstereography of the spine as a diagnostic test for adolescent idiopathic soliosis (AIS), and to compare its results with those obtained using a scoliometer. Methods. Adolescents suspected of AIS and scheduled for radiographs were included. Rasterstereographic scoliosis angle (SA), maximal vertebral surface rotation (ROT), and angle of trunk rotation (ATR) with a scoliometer were evaluated. The area under the curve (AUC) from receiver operating characteristic (ROC) plots were used to describe the discriminative ability of the SA, ROT, and ATR for scoliosis, defined as a Cobb angle > 10°. Test characteristics (sensitivity and specificity) were reported for the best threshold identified using the Youden method. AUC of SA, ATR, and ROT were compared using the bootstrap test for two correlated ROC curves method. Results. Of 212 patients studied, 146 (69%) had an AIS. The AUC was 0.74 for scoliosis angle (threshold 12.5°, sensitivity 75%, specificity 65%), 0.65 for maximal vertebral surface rotation (threshold 7.5°, sensitivity 63%, specificity 64%), and 0.82 for angle of trunk rotation (threshold 5.5°, sensitivity 65%, specificity 80%). The AUC of ROT was significantly lower than that of ATR (p < 0.001) and SA (p < 0.001). The AUCs of ATR and SA were not significantly different (p = 0.115). Conclusion. The rasterstereographic scoliosis angle has better diagnostic characteristics than the angle of trunk rotation evaluated with a scoliometer, with similar AUCs and a higher sensitivity. Cite this article: Bone Joint J 2023;105-B(4):431–438

Aims. To compare the cost-effectiveness of high-dose, dual-antibiotic cement versus single-antibiotic cement for the treatment of displaced intracapsular hip fractures in older adults. Methods. Using data from a multicentre randomized controlled trial (World Hip Trauma Evaluation 8 (WHiTE-8)) in the UK, a within-trial economic evaluation was conducted. Resource usage was measured over 120 days post randomization, and cost-effectiveness was reported in terms of incremental cost per quality-adjusted life year (QALY), gained from the UK NHS and personal social services (PSS) perspective in the base-case analysis. Methodological uncertainty was addressed using sensitivity analysis, while decision uncertainty was handled using confidence ellipses and cost-effectiveness acceptability curves. Results. The base-case analysis showed that high-dose, dual-antibiotic cement had a significantly higher mean cost (£224 (95% confidence interval (CI) -408 to 855)) and almost the same QALYs (0.001 (95% CI -0.002 to 0.003)) relative to single-antibiotic cement from the UK NHS and PSS perspective. The probability of the high-dose, dual-antibiotic cement being cost-effective was less than 0.3 at alternative cost-effectiveness thresholds, and its net monetary benefit was negative. This finding remained robust in the sensitivity analyses. Conclusion. This study shows that high-dose, dual-antibiotic cement is unlikely to be cost-effective compared to single-antibiotic cement for the treatment of displaced intracapsular hip fractures in older adults. Cite this article: Bone Joint J 2023;105-B(10):1070–1077

Aims. The aim of this study was to compare the cost-effectiveness of surgical fixation with Kirschner (K-)wire ersus moulded casting after manipulation of a fracture of the distal radius in an operating theatre setting. Methods. An economic evaluation was conducted based on data collected from the Distal Radius Acute Fracture Fixation Trial 2 (DRAFFT2) multicentre randomized controlled trial in the UK. Resource use was collected at three, six, and 12 months post-randomization using trial case report forms and participant-completed questionnaires. Cost-effectiveness was reported in terms of incremental cost per quality-adjusted life year (QALY) gained from an NHS and personal social services perspective. Sensitivity analyses were conducted to examine the robustness of cost-effectiveness estimates, and decision uncertainty was handled using confidence ellipses and cost-effectiveness acceptability curves. Results. In the base case analysis, surgical fixation with K-wire was more expensive (£29.65 (95% confidence interval (CI) -94.85 to 154.15)) and generated lower QALYs (0.007 (95% CI -0.03 to 0.016)) than moulded casting, but this difference was not statistically significant. The probability of K-wire being cost-effective at a £20,000 per QALY cost-effectiveness threshold was 24%. The cost-effectiveness results remained robust in the sensitivity analyses. Conclusion. The findings suggest that surgical fixation with K-wire is unlikely to be a cost-effective alternative to a moulded cast in adults, following manipulation of a fracture of the distal radius in a theatre setting. Cite this article: Bone Joint J 2022;104-B(11):1225–1233

Aims. Clinical management of open fractures is challenging and frequently requires complex reconstruction procedures. The Gustilo-Anderson classification lacks uniform interpretation, has poor interobserver reliability, and fails to account for injuries to musculotendinous units and bone. The Ganga Hospital Open Injury Severity Score (GHOISS) was designed to address these concerns. The major aim of this review was to ascertain the evidence available on accuracy of the GHOISS in predicting successful limb salvage in patients with mangled limbs. Methods. We searched electronic data bases including PubMed, CENTRAL, EMBASE, CINAHL, Scopus, and Web of Science to identify studies that employed the GHOISS risk tool in managing complex limb injuries published from April 2006, when the score was introduced, until April 2021. Primary outcome was the measured sensitivity and specificity of the GHOISS risk tool for predicting amputation at a specified threshold score. Secondary outcomes included length of stay, need for plastic surgery, deep infection rate, time to fracture union, and functional outcome measures. Diagnostic test accuracy meta-analysis was performed using a random effects bivariate binomial model. Results. We identified 1,304 records, of which six prospective cohort studies and two retrospective cohort studies evaluating a total of 788 patients were deemed eligible for inclusion. A diagnostic test meta-analysis conducted on five cohort studies, with 474 participants, showed that GHOISS at a threshold score of 14 has a pooled sensitivity of 93.4% (95% confidence interval (CI) 78.4 to 98.2) and a specificity of 95% (95% CI 88.7 to 97.9) for predicting primary or secondary amputations in people with complex lower limb injuries. Conclusion. GHOISS is highly accurate in predicting success of limb salvage, and can inform management and predict secondary outcomes. However, there is a need for high-quality multicentre trials to confirm these findings and investigate the effectiveness of the score in children, and in predicting secondary amputations. Cite this article: Bone Joint J 2023;105-B(1):21–28

Aims. Histology is widely used for diagnosis of persistent infection during reimplantation in two-stage revision hip and knee arthroplasty, although data on its utility remain scarce. Therefore, this study aims to assess the predictive value of permanent sections at reimplantation in relation to reinfection risk, and to compare results of permanent and frozen sections. Methods. We retrospectively collected data from 226 patients (90 hips, 136 knees) with periprosthetic joint infection who underwent two-stage revision between August 2011 and September 2021, with a minimum follow-up of one year. Histology was assessed via the SLIM classification. First, we analyzed whether patients with positive permanent sections at reimplantation had higher reinfection rates than patients with negative histology. Further, we compared permanent and frozen section results, and assessed the influence of anatomical regions (knee versus hip), low- versus high-grade infections, as well as first revision versus multiple prior revisions on the histological result at reimplantation. Sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), chi-squared tests, and Kaplan-Meier estimates were calculated. Results. Overall, the reinfection rate was 18%. A total of 14 out of 82 patients (17%) with positive permanent sections at reimplantation experienced reinfection, compared to 26 of 144 patients (18%) with negative results (p = 0.996). Neither permanent sections nor fresh frozen sections were significantly associated with reinfection, with a sensitivity of 0.35, specificity of 0.63, PPV of 0.17, NPV of 0.81, and accuracy of 58%. Histology was not significantly associated with reinfection or survival time for any of the analyzed sub-groups. Permanent and frozen section results were in agreement for 91% of cases. Conclusion. Permanent and fresh frozen sections at reimplantation in two-stage revision do not serve as a reliable predictor for reinfection. Cite this article: Bone Joint J 2024;106-B(4):372–379

Aims. The early diagnosis of cauda equina syndrome (CES) is crucial for a favourable outcome. Several studies have reported the use of an ultrasound scan of the bladder as an adjunct to assess the minimum post-void residual volume of urine (mPVR). However, variable mPVR values have been proposed as a threshold without consensus on a value for predicting CES among patients with relevant symptoms and signs. The aim of this study was to perform a meta-analysis and systematic review of the published evidence to identify a threshold mPVR value which would provide the highest diagnostic accuracy in patients in whom the diagnosis of CES is suspected. Methods. The search strategy used electronic databases (PubMed, Medline, EMBASE, and AMED) for publications between January 1996 and November 2021. All studies that reported mPVR in patients in whom the diagnosis of CES was suspected, followed by MRI, were included. Results. A total of 2,115 studies were retrieved from the search. Seven fulfilled the inclusion criteria. These included 1,083 patients, with data available from 734 being available for meta-analysis. In 125 patients, CES was confirmed by MRI. The threshold value of mPVR reported in each study varied and could be categorized into 100 ml, 200 ml, 300 ml, and 500 ml. From the meta-analysis, 200 ml had the highest diagnostic accuracy, with 82% sensitivity (95% confidence interval (CI) 0.72 to 0.90) and 65% specificity (95% CI 0.70 to 0.90). When compared using summative receiver operating characteristic curves, mPVR of 200 ml was superior to other values in predicting the radiological confirmation of CES. Conclusion. mPVR is a useful tool when assessing patients in whom the diagnosis of CES is suspected. Compared with other values a mPVR of 200 ml had superior sensitivity, specificity, and positive and negative predictive values. In a patient with a suggestive history and clinical findings, a mPVR of > 200 ml should further raise the suspicion of CES. Caution is recommended when considering the mPVR in isolation and using it as an ‘exclusion tool’, and it should only be used as an adjunct to a full clinical assessment. Cite this article: Bone Joint J 2023;105-B(5):487–495

Aims. The aim of this study was to evaluate the optimal deep tissue specimen sample number for histopathological analysis in the diagnosis of periprosthetic joint infection (PJI). Methods. In this retrospective diagnostic study, patients undergoing revision surgery after total hip or knee arthroplasty (n = 119) between January 2015 and July 2018 were included. Multiple specimens of the periprosthetic membrane and pseudocapsule were obtained for histopathological analysis at revision arthroplasty. Based on the Infectious Diseases Society of America (IDSA) 2013 criteria, the International Consensus Meeting (ICM) 2018 criteria, and the European Bone and Joint Infection Society (EBJIS) 2021 criteria, PJI was defined. Using a mixed effects logistic regression model, the sensitivity and specificity of the histological diagnosis were calculated. The optimal number of periprosthetic tissue specimens for histopathological analysis was determined by applying the Youden index. Results. Based on the EBJIS criteria (excluding histology), 46 (39%) patients were classified as infected. Four to six specimens showed the highest Youden index (four specimens: 0.631; five: 0.634; six: 0.632). The sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) of five tissue specimens were 76.5% (95% confidence interval (CI) 67.6 to 81.4), 86.8% (95% CI 81.3 to 93.5), 66.0% (95% CI 53.2 to 78.7), and 84.3% (95% CI 79.4 to 89.3), respectively. The area under the curve (AUC) was calculated with 0.81 (as a function of the number of tissue specimens). Applying the ICM and IDSA criteria (excluding histology), 40 (34%) and 32 (27%) patients were categorized as septic. Three to five specimens had the highest Youden index (ICM 3: 0.648; 4: 0.651; 5: 0.649) (IDSA 3: 0.627; 4: 0.629; 5: 0.625). Conclusion. Three to six tissue specimens of the periprosthetic membrane and pseudocapsule should be collected at revision arthroplasty and analyzed by a pathologist experienced and skilled in interpreting periprosthetic tissue. Cite this article: Bone Joint J 2023;105-B(2):158–165

Aims. The preoperative diagnosis of periprosthetic joint infection (PJI) remains a challenge due to a lack of biomarkers that are both sensitive and specific. We investigated the performance characteristics of polymerase chain reaction (PCR), interleukin-6 (IL6), and calprotectin of synovial fluid in the diagnosis of PJI. Methods. We performed systematic search of PubMed, Embase, The Cochrane Library, Web of Science, and Science Direct from the date of inception of each database through to 31 May 2021. Studies which described the diagnostic accuracy of synovial fluid PCR, IL6, and calprotectin using the Musculoskeletal Infection Society criteria as the reference standard were identified. Results. Overall, 31 studies were identified: 20 described PCR, six described IL6, and five calprotectin. The sensitivity and specificity were 0.78 (95% confidence interval (CI) 0.67 to 0.86) and 0.97 (95% CI 0.94 to 0.99), respectively, for synovial PCR;, 0.86 (95% CI 0.74 to 0.92), and 0.94 (95% CI 0.90 to 0.96), respectively, for synovial IL6; and 0.94 (95% CI 0.82 to 0.98) and 0.93 (95% CI 0.85 to 0.97), respectively, for synovial calprotectin. Likelihood ratio scattergram analyses recommended clinical utility of synovial fluid PCR and IL6 as a confirmatory test only. Synovial calprotectin had utility in the exclusion and confirmation of PJI. Conclusion. Synovial fluid PCR and IL6 had low sensitivity and high specificity in the diagnosis of PJI, and is recommended to be used as confirmatory test. In contrast, synovial fluid calprotectin had both high sensitivity and specificity with utility in both the exclusion and confirmation of PJI. We recommend use of synovial fluid calprotectin studies in the preoperative workup of PJI. Cite this article: Bone Joint J 2022;104-B(3):311–320

Aims. Gram-negative periprosthetic joint infection (PJI) has been poorly studied despite its rapidly increasing incidence. Treatment with one-stage revision using intra-articular (IA) infusion of antibiotics may offer a reasonable alternative with a distinct advantage of providing a means of delivering the drug in high concentrations. Carbapenems are regarded as the last line of defense against severe Gram-negative or polymicrobial infection. This study presents the results of one-stage revision using intra-articular carbapenem infusion for treating Gram-negative PJI, and analyzes the characteristics of bacteria distribution and drug sensitivity. Methods. We retrospectively reviewed 32 patients (22 hips and 11 knees) who underwent single-stage revision combined with IA carbapenem infusion between November 2013 and March 2020. The IA and intravenous (IV) carbapenem infusions were administered for a single Gram-negative infection, and IV vancomycin combined with IA carbapenems and vancomycin was applied for polymicrobial infection including Gram-negative bacteria. The bacterial community distribution, drug sensitivity, infection control rate, functional recovery, and complications were evaluated. Reinfection or death caused by PJI was regarded as a treatment failure. Results. Gram-negative PJI was mainly caused by Escherichia coli (8/34), Enterobacter cloacae (7/34), and Klebsiella pneumoniae (5/34). Seven cases (7/32) involved polymicrobial PJIs. The resistance rates of penicillin, cephalosporin, quinolones, and sulfonamides were > 10%, and all penicillin and partial cephalosporins (first and second generation) were > 30%. Of 32 cases, treatment failed to eradicate infection in only three cases (9.4%), at a mean follow-up of 55.1 months (SD 25 to 90). The mean postoperative Harris Hip Score and Hospital for Special Surgery knee score at the most recent follow-up were 81 (62 to 91) and 79 (56 to 89), respectively. One patient developed a fistula, and another presented with a local rash on an infected joint. Conclusion. The use of IA carbapenem delivered alongside one-stage revision effectively controlled Gram-negative infection and obtained acceptable clinical outcomes with few complications. Notably, first- and second-generation cephalosporins and penicillin should be administrated with caution, due to a high incidence of resistance. Cite this article: Bone Joint J 2023;105-B(3):284–293