Hallucination Rates and Reference Accuracy of ChatGPT and Bard for Systematic Reviews: Comparative Analysis.

BACKGROUND: Large language models (LLMs) have raised both interest and concern in the academic community. They offer the potential for automating literature search and synthesis for systematic reviews but raise concerns regarding their reliability, as the tendency to generate unsupported (hallucinated) content persist. OBJECTIVE: The aim of the study is to assess the performance of LLMs such as ChatGPT and Bard (subsequently rebranded Gemini) to produce references in the context of scientific writing. METHODS: The performance of ChatGPT and Bard in replicating the results of human-conducted systematic reviews was assessed. Using systematic reviews pertaining to shoulder rotator cuff pathology, these LLMs were tested by providing the same inclusion criteria and comparing the results with original systematic review references, serving as gold standards. The study used 3 key performance metrics: recall, precision, and F1-score, alongside the hallucination rate. Papers were considered "hallucinated" if any 2 of the following information were wrong: title, first author, or year of publication. RESULTS: In total, 11 systematic reviews across 4 fields yielded 33 prompts to LLMs (3 LLMs×11 reviews), with 471 references analyzed. Precision rates for GPT-3.5, GPT-4, and Bard were 9.4% (13/139), 13.4% (16/119), and 0% (0/104) respectively (P<.001). Recall rates were 11.9% (13/109) for GPT-3.5 and 13.7% (15/109) for GPT-4, with Bard failing to retrieve any relevant papers (P<.001). Hallucination rates stood at 39.6% (55/139) for GPT-3.5, 28.6% (34/119) for GPT-4, and 91.4% (95/104) for Bard (P<.001). Further analysis of nonhallucinated papers retrieved by GPT models revealed significant differences in identifying various criteria, such as randomized studies, participant criteria, and intervention criteria. The study also noted the geographical and open-access biases in the papers retrieved by the LLMs. CONCLUSIONS: Given their current performance, it is not recommended for LLMs to be deployed as the primary or exclusive tool for conducting systematic reviews. Any references generated by such models warrant thorough validation by researchers. The high occurrence of hallucinations in LLMs highlights the necessity for refining their training and functionality before confidently using them for rigorous academic purposes.

Subjective Shoulder Value for Sport Is a Simple, Reliable, and Valid Score to Assess Shoulder Function in Athletes.

PURPOSE: To validate the subjective shoulder value for sport (SSV-Sport) by measuring its correlation with existing patient-reported outcome measures (PROMs) and defining its psychometric properties. METHODS: Between May 2021 and May 2022, we established 2 patient groups. Group 1 included those (1) aged 18 years or older, who were (2) consulting for the first time for any shoulder condition, (3) regularly participated in sports, and were capable of accessing a questionnaire independently. There were asked to rate their SSV and SSV-Sport at admission and 2 weeks later; they also were asked to answer a questionnaire including other PROMS. Group 2 comprised patients who had (1) undergone shoulder stabilization surgery and had (2) a minimum follow-up period of 6 months. RESULTS: For the shoulder disability patients (group 1, n = 62), there was a strong and significant correlation between SSV-Sport and other PROMs: Quick Disabilities of the Arm, Shoulder and Hand Sport (r = 0.84), Walch-Duplay (r = 0.65), Rowe (r = 0.74), Western Ontario Shoulder Instability (r = 0.78), and SSV (r = 0.75) (P = .0001). The SSV-Sport was reliable at baseline and 2 weeks after (0.91, 95% confidence interval 0.85-0.94), and was responsive to change (P < .001). For the anterior instability patients (group 2, n = 83), SSV was on average 50 points greater than SSV-Sport (29.2 vs 79.4, P < .001) for preoperative values. In both groups, the values of SSV were constantly and significantly higher than the values of SSV-Sport (81.9 ± 21.3 vs 54.8 ± 30.9; P < .001). CONCLUSIONS: The SSV-Sport is an easily administered, reliable, responsive, and valid measure of shoulder function in athletes that is highly correlated with other PROMs. SSV-Sport is better adapted than SSV to quantify pre- and postoperative shoulder deficiency in athletes. LEVEL OF EVIDENCE: Level III, cohort study (diagnosis).

Glenoid lateralization and subscapularis repair are independent predictive factors of improved internal rotation after reverse shoulder arthroplasty.

PURPOSE: Reverse shoulder arthroplasty (RSA) has shown improvement in clinical outcomes for various conditions, although some authors expressed concern about the restoration of active internal rotation (AIR). The current study assesses preoperative and intraoperative predictive factors of AIR in patients having a Grammont-style RSA with a minimum five year follow-up. METHODS: We conducted a retrospective multicentric study, including patients operated on with a 155° Grammont-style RSA for cuff-related pathology or primary osteoarthritis with posterior subluxation or an associated cuff tear. Patients were clinically evaluated at a minimum of five year follow-up. Patients with previous surgery or those who had a tendon transfer with the RSA were excluded. Demographic parameters, BMI, preoperative notes, and operative reports were obtained from medical records. AIR was graded according to the constant score system from 0 to 10. RESULTS: A total of 280 shoulders in 269 patients (mean age at surgery, 74.9 ± 5.9 years) met the inclusion criteria and were analyzed. The average follow-up was 8.1 years (range, 5-16 years). Overall, AIR increased from 4.2 (SD 2.5, range 0 to 10) preoperatively to 5.9 (SD 2.6, range 0 to 10) at final follow-up. At the last follow-up, AIR increased in 56% of cases, was unchanged in 26% and decreased in 18%. In 188 shoulders (67%), internal rotation was functional and allowed patients to reach the level of L3 or higher. Multivariable linear regression found the following preoperative clinical factors predictive of worse AIR after RSA: male gender (ß = -1.25 [-2.10; -0.40]; p = 0.0042) and higher values of BMI (ß = -0.085 [-0.17; -0.0065]; p = 0.048). Two surgical factors were associated with better AIR after RSA: glenoid lateralization with BIO-RSA technique (ß = 0.80 [0.043; 1.56]; p = 0.039) and subscapularis repair (ß = 1.16 [0.29; 2.02]; p = 0.0092). CONCLUSIONS: With a mean of eight year follow-up (5 to 16 years), internal rotation was functional (≥ L3 level) in 67% of operated shoulders after Grammont-style RSA; however, two patients out of ten had decreased AIR after surgery. Male patients and those with higher BMIs had worse AIR, with glenoid lateralization (using the BIO-RSA technique) and subscapularis repair, as they are predictive of increased AIR after RSA. LEVEL OF EVIDENCE: Case series, Level IV.

Screw-Related Complications May Occur at a Greater Rate After Arthroscopic Versus Open Latarjet Procedure: A Systematic Review.

Purpose: To determine the rate of complications attributable to the use of screw fixation during the Latarjet procedure and to delineate screw-related complications for open and arthroscopic Latarjet surgery. Methods: A systematic review of the literature was performed by querying MEDLINE and EMBASE computerized databases for relevant articles that reported clinical outcomes associated with the Latarjet procedure. Clinical studies of open and/or arthroscopic Latarjet surgery that employed screw fixation of the coracoid were included in our analysis. Results: From 692 articles identified initially, 32 studies met eligibility criteria. The study cohort comprised 2,758 shoulders, with a mean age of patients ranged from 17 to 62 years, and the mean duration of follow-up ranged from 0.3 to 25.6 years. Twenty-two studies reported outcomes of an open Latarjet technique whereas 10 studies reported on an arthroscopic Latarjet technique. Across all studies, the overall complication rate ranged from 1.4% to 36%. The rate of screw-related complications ranged from 0% to16%, and the rate of screw removal ranged from 0% to 18%. Among the subset of studies that reported specific indications for screw removal, the most common indications were pain and screw loosening. Reported rates of screw-removal among arthroscopic Latarjet procedures range from 0% to 18% and among open procedures range from 0% to 7.3%. Conclusions: (1) Up to one-third of the overall surgical complications associated with the Latarjet procedure may be related to the use of screw fixation, (2) arthroscopic Latarjet procedure did not decrease the rate of screw related complications. The reported rate of screw-removal may be higher after the arthroscopic Latarjet procedures (up to 18%) than after open procedures (up to 7,3%), and is mainly indicated for persistent shoulder pain and/or screw loosening. Level of Evidence: Level IV, systematic review of Level III-IV studies.

Short term results of anatomic stemless shoulder replacement with peripheral neck fixation.

Background: Most of the current stemless implants are using a central neck fixation design. Easytech Stemless replacements (FX Solutions®, Viriat, France) were developed promoting the idea of peripheral fixation of the stemless implant. The aim of this study was to analyze clinical and radiological results of this stemless TSA implanted for osteoarthritis (OA). Methods: A retrospective study of patients who received TSA with the Easytech® Anatomical Shoulder System (FX Solutions®, Viriat, France) for OA was completed at five clinical sites. Preoperative and minimum 2 year follow-up Constant Scores and X-rays were evaluated. Two independent orthopedic surgeons analyzed X-rays to assess anatomy reconstruction and component migration. Revisions and serious adverse effects were reviewed. Results: 129 patients were analyzed with an average follow-up of 37.47 months (24-54). Seven patients had a revision surgery and one of these resulted in the removal of the Anchor Base. The postoperative Constant Scores showed significant improvement with a mean raw Constant Score of 73.78 (38.0-100.0, mean increase 40.96, P < .0001) and mean adjusted Constant Score of 99.14 (42.20-133.30, mean increase 55.68, P < .0001). Radiographic review demonstrated that no modification of the center of rotation (COR) at minimum 24 months was greater than 3 mm and in comparison to preoperative radiographs, restoration of COR was lateralized in 25% of the cases, medialized in 61.5% and 13.5% had no change in COR. The mean value of the cervico-diaphyseal angle was 130.2° (114; 149) at 3 weeks and 129.9° (113.5; 144.0) at 2 years, demonstrating no significant difference (P = .16). 36 patients (35%) had calcar remodeling. Univariate and multivariate logistic regression found no significant factor, which was associated with the variation of the cervico-diaphyseal angle over time, center of rotation, calcar remodeling, or final Constant Score. The Walch B2 and C glenoid subset of patients (n = 13) had overall survival rate of 90.9% [95% CI 75.4-100.0] vs. 98.0% [95% CI 95.3-100.0] in other glenoid types (P < .01). Walch B2/C patients were 4.44 [95% CI 1.13, 17.6] times more likely to have a Constant Score <85 (P = .03). Conclusions: The peripheral fixation of the Easytech® Anatomical Shoulder System for OA provides excellent clinical results and imaging stability at minimum two year follow-up. Elderly age, gender and BMI do not affect the stability.

Stemless reverse shoulder arthroplasty: clinical and radiologic outcomes with minimum 2 years' follow-up.

BACKGROUND: Recently, a stemless reverse shoulder arthroplasty (RSA) design was developed to preserve bone stock. Clinical and radiologic studies of this design in larger cohorts with >100 patients are not frequent. The purpose of this study was to present the clinical and radiologic results of a newly developed stemless RSA implant. The hypothesis was that this design would provide similar clinical and radiologic results to other stemless implants, as well as stemmed implants. METHODS: Between September 2015 and December 2019, all patients who underwent primary RSA with a stemless Easytech prosthesis were considered eligible for inclusion in this prospective multicenter study. The minimum follow-up period was 2 years. Clinical outcomes consisted of the Constant score, adjusted Constant score, QuickDASH (short version of Disabilities of the Arm, Shoulder and Hand questionnaire) score, Subjective Shoulder Value, and American Shoulder and Elbow Surgeons shoulder score. Radiographic parameters included radiolucency, loosening, scapular notching, and specific geometric parameters. RESULTS: Stemless RSA was performed in 115 patients (61 women and 54 men) at 6 different clinical centers. The average age at the time of surgery was 68.7 years. The average Constant score was 32.5 preoperatively and showed significant improvement to 61.8 at latest follow-up (P < .001). The Subjective Shoulder Value also demonstrated significant improvement postoperatively (from 27.0 to 77.5, P < .001). Scapular notching was observed in 28 patients (24.3%); humeral loosening, 5 (4.3%); and glenoid loosening, 4 (3.5%). The total complication rate was 17.4%. Eight patients (4 women and 4 men) underwent implant revision. CONCLUSION: The clinical outcomes of the examined stemless RSA seem to be comparable to those of other humeral designs; however, the complication and revision rates are higher than those of historical controls. Surgeons should proceed with caution when using this implant until longer-term follow-up data are available.

Arthroscopic revision cuff repair: do tendons have a second chance to heal?

BACKGROUND: Few studies have investigated postoperative tendon integrity after reoperation for failed rotator cuff repair. The purpose of this study was to evaluate the anatomic and clinical outcomes of arthroscopic revision rotator cuff repair (AR-RCR) and identify the risk factors related to re-retear. METHODS: Sixty-nine consecutive patients (mean age, 55 years) with primary failed open (38%) or arthroscopic (62%) cuff repairs underwent AR-RCR and were reviewed regarding clinical examination findings and imaging studies. Patients with massive cuff tears and upward humeral migration (acromiohumeral distance < 6 mm) or glenohumeral osteoarthritis were excluded. Revision repair was performed by a single, experienced shoulder surgeon. Complete footprint coverage was achieved in all cases using a single-row (70%), double-row (19%), or side-to-side (11%) technique. The primary outcome measure was tendon healing assessed with magnetic resonance imaging (57 cases) or computed tomography arthrogram (12 cases) performed at minimum 1-year follow-up. Secondary outcome measures included functional outcome scores, subjective results, and complications. The mean follow-up period was 43 months (range, 12-136 months). RESULTS: The cuff tendons did not heal to the tuberosity in 36% of the shoulders (25 of 69) following revision cuff surgery. Absence of tendon healing was associated with poorer shoulder function (average Constant score, 69 ± 20 vs. 54 ± 18; P = .003) and a decreased Subjective Shoulder Value (72% vs. 54%, P = .002). Factors that were negatively associated with tendon healing were age ≥ 55 years (odds ratio [OR], 4.5 [95% confidence interval, 1.6-12.5]; P = .02), tendon retraction of stage 2 or higher (OR, 4.4 [95% confidence interval, 1.4-14.3]; P = .01), and fatty infiltration index > 2 (OR, 10.2; P < .0001). No differences in retear rates were found between single-row and double-row cases. In 36 shoulders, tissue samples were harvested and submitted for bacteriologic culture analysis; 13 (36%) showed positive findings for infection (Cutibacterium acnes in 12 of 13) and associated antibiotic treatment was given. Overall, 25% of patients had unsatisfactory clinical results and 22% were disappointed or dissatisfied. At last follow-up, 4 patients (5.7%) underwent reoperations, with a second AR-RCR in 1 and conversion to reverse shoulder arthroplasty in 3. CONCLUSION: Despite careful patient selection and intraoperative complete footprint coverage, in this study the tendons did not heal to bone in 36% of cases after revision cuff surgery. The absence of tendon healing is associated with poorer clinical and subjective results. Patients aged ≥ 55 years and patients with larger tears (stage 2 or higher) and/or muscle fatty infiltration (fatty infiltration index > 2) have significantly lower rates of healing. Surgeons should be aware that structurally failed cuff repair may also be associated with low-grade infection.

Idiopathic endogenous lipid pneumonia: A case report of a young Syrian man.

Lipid pneumonia is pneumonia due to aspiration or inhalation of various oily or fatty substances. It can be divided into exogenous and endogenous depending on the source of the lipids. In endogenous lipid pneumonia (ELP), lipid accumulates in the intra alveoli as a result of obstruction, chronic lung infection/disease, or a lipid storage disorder. This study presents a case of a 47 year-old man with a history of smoking, surgically repaired ventricular septal defect, pulmonary stenosis, and no history of lipid intake. He complained of worsened exertional dyspnea and a chronic non-productive cough with no signs and symptoms of infection. The patient was diagnosed with idiopathic endogenous lipid pneumonia after excluding any inflammatory or systemic disease as a possible cause. Open lung biopsy showed lipid-laden macrophages. Corticosteroids were the mainstay therapy with no radiological improvement, eventually death occurred due to pulmonary embolism.

Outpatient shoulder prostheses: Feasibility, acceptance and safety.

INTRODUCTION: Outpatient surgery in France is defined by the national authority for health (HAS) as a scheduled surgery enabling same-day discharge without any increased risk to the patient. With the advent of enhanced recovery after surgery, outpatient lower limb arthroplasty has become a common procedure. However, only 1.1% of knee arthroplasties in France were performed on an outpatient basis in 2017. OBJECTIVES: 1) assess early morbidity and mortality after outpatient shoulder arthroplasties to validate eligibility and safety criteria; and 2) assess patient acceptance of outpatient surgery. METHODS: A single-center study with the following inclusion criteria: primary shoulder arthroplasty, American Society of Anesthesiology (ASA) score I or II, no cognitive impairment, and no coronary artery or thromboembolic diseases. Analgesia was provided by bupivacaine via a peripheral nerve catheter in the first 72 hours followed by oral analgesics. Patients were discharged if the post-anesthetic discharge scoring system (PADSS) was>9/10 and the visual analog scale (VAS) was<5/10. Postoperative telephone interviews were carried out on D1, D2 and D3 to assess pain with the numerical rating scale and to collect data on their analgesic consumption. All patients were seen by an independent observer at one and six months for a clinical and radiologic follow-up and at 90 days during a consultation with the senior surgeon. The primary endpoint was the 90-day morbidity and mortality rate (readmissions, rehospitalizations, and minor and major complications). A satisfaction questionnaire was collected at one and six months. RESULTS: Thirty-six patients were offered an outpatient shoulder arthroplasty between February 2016 and February 2018: 12 (33%) refused with no valid reasons and 24 patients agreed to the procedure (seven hemiarthroplasties, nine anatomic shoulder arthroplasties and eight reverse shoulder arthroplasties). The mean age at surgery was 70 years (55-82), mean body mass index (BMI) was 26 (21-32) and 14 patients were ASA II (66%). Three patients (12%) refused same-day discharge despite a PADSS score>9/10 and adequate pain management. Two patients (8%) were not discharged home on the same day as the surgery for medical reasons (one for pain and one for high blood pressure). No readmissions or complications were reported for the 19 outpatient arthroplasties. None of the outpatients used opioids. All patients were satisfied with their functional outcome, 84% were satisfied with the outpatient management and 17% felt they were insufficiently monitored and regretted that they were not hospitalized. CONCLUSIONS: 1) outpatient shoulder arthroplasty can be safely proposed to selected patients with low comorbidities, regardless of their age and type of implant; 2) the acceptance rate for outpatient shoulder arthroplasty remained low among our patient population. These results should incite us to better educate patients about outpatient surgery. LEVEL OF EVIDENCE: IV; retrospective study.

One-session bilateral total knee replacement: Late complications and survivorship.

INTRODUCTION: The early morbidity and mortality of one-session bilateral total knee arthroplasty (1-session BTKA) has been reported in the medical literature. However, there is less information about the long-term clinical impact of this strategy. The aim of this study was to report on the late complications (>90 days), clinical outcomes (KOOS and new KSS) along with the survivorship of 1-session BTKA. We hypothesised that 1-session BTKA will cause few late complications and that the implant survival will meet the criteria of the National Institute for Health and Clinical Excellence (NICE). METHODS: This single-centre retrospective study analysed a cohort of ASA-1 and ASA-2 patients who underwent 1-session BTKR over an 8-year period (2009 to 2016). The cohort consisted of 116 patients (66% women, 34% men) with mean age at inclusion of 69 years (32-85 years); 22% of patients were ASA-1 and 78% were ASA-2. The implant-related, infection-related or mechanical complications that occurred more than 90 days after the operation, the clinical outcomes (KOOS and New KSS) and radiological outcomes, along with the survivorship were determined during the scheduled follow-up visits at 3 months, 6 months and 1 year postoperative and during the long-term follow-up or during an intercurrent event. RESULTS: No patients were lost to follow-up. Fifteen complications occurred (6.5%): four infections, four patellar problems (three cases of clunk syndrome and one of patellofemoral pain), four cases of stiffness, two of unexplained pain and one femoral periprosthetic fracture. Eight patients were readmitted to the hospital (7%); seven were reoperated (3%) and two implants were revised (1%). The functional outcomes (KOOS and New KSS) were significantly improved and 87% of patients were satisfied or very satisfied with this procedure. At a mean follow-up of 5 years, the survivorship estimated using the Kaplan-Meier method was 98.4% (95% CI: 0.933-0.996). CONCLUSION: Performing 1-session BTKA is a reliable strategy as it produces a low rate of late complications, excellent medium-term functional outcomes and survivorship that meets NICE criteria, thus confirming our hypothesis. LEVEL OF EVIDENCE: IV, retrospective case series.

Early morbidity and mortality after single-stage bilateral total knee replacement.

INTRODUCTION: Single-stage bilateral total knee replacement (TKR) has the advantages of requiring only one hospital stay and one anesthesia session, having a shorter rehabilitation period, and reducing the cost of patient care. However, this strategy is controversial because of the perioperative risk. We hypothesized that this strategy did not cause early perioperative mortality and that the early morbidity and readmission rates would be low when patients are selected based on their ASA score. METHODS: This single-center retrospective study analyzed a cohort of ASA-1 and ASA-2 patients who underwent single-stage bilateral TKR over an 8-year period (2009 to 2016). The study cohort consisted of 116 patients, mainly women with mean age of 69 years at inclusion; 22.4% of patients were ASA-1 and 77.6% were ASA-2. Death and early complications during the first 90 days postoperative, the early readmission rate and the blood-sparing strategy were analyzed using the clinical and paraclinical data collected during the hospital stay, during the convalescent care center stay, and during the follow-up visits at 6 weeks and 3 months postoperative. The analysis was completed using the intrahospital software Clinicom, which allowed us to trace all the events and episodes for each patient. RESULTS: The early mortality rate was 0%. There were five major complications (4.3%) and thirteen minor complications (11%). The early readmission rate was 5.2%. Homologous blood transfusion was performed in 36% of patients. Administration of tranexamic acid reduced this rate to 24.3% versus 44% in patients not taking it (p=0.06). CONCLUSION: The perioperative mortality in this selected population is zero and the early morbidity is acceptable. The early readmission rate is also low. Thus proposing single-stage bilateral TKR to patients meeting the criteria defined in this study is a valid strategy. LEVEL OF EVIDENCE: IV, retrospective cohort study.