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.

Avoiding superior inclination of the glenoid component in reverse shoulder arthroplasty (RSA) is crucial. We hypothesized that superior inclination was underestimated in RSA. Our purpose was to describe and assess a new measurement of inclination for the inferior portion of the glenoid (where the baseplate rests). The study included 47 shoulders with rotator cuff tear arthropathy (mean age, 76 years). The reverse shoulder arthroplasty angle (RSA angle), defined as the angle between the inferior part of the glenoid fossa and the perpendicular to the floor of the supraspinatus, was compared with the global glenoid inclination (β angle or total shoulder arthroplasty [TSA] angle). Measurements were made on plain anteroposterior radiographs and reformatted 2-dimensional (2D) computed tomography (CT) scans by 3 independent observers and compared with 3-dimensional (3D) software (Glenosys) measurements. The mean RSA angle was 25° ± 8° on plain radiographs, 20° ± 6° on reformatted 2D CT scans, and 21° ± 5° via 3D reconstruction software. The mean TSA angle was on average 10° ± 5° lower than the mean RSA angle (P < .001); this difference was observed regardless of the method of measurement (radiographs, 2D CT, or 3D CT) and type of glenoid erosion according to Favard. In Favard type E1 glenoids with central concentric erosion, the difference between the 2 angles was 12° ± 4° (P < .001). The same angle cannot be used to measure glenoid inclination in anatomic and reverse prostheses. The TSA (or β) angle underestimates the superior orientation of the reverse baseplate in RSA. The RSA angle (20° ± 5°) needs to be corrected to achieve neutral inclination of the baseplate (RSA angle = 0°). Surgeons should be aware that E1 glenoids (with central erosion) are at risk for baseplate superior tilt if the RSA angle is not corrected.

Reverse Shoulder Arthroplasty (RSA) may be indicated in displaced proximal humerus fractures in elderly patients. We hypothesized that tuberosity fixation and healing around the prosthesis would result in better outcomes and patient satisfaction. Thirty-eight acute displaced or dislocated 3- and 4-part fractures in elderly patients were treated with reattachment of the tuberosities around a RSA. The mean age at surgery was 80 ± 4 years (range, 70-88 years). A specific reverse fracture stem that incorporated a cancellous bone autograft (harvested from the fractured head) and a standardized suturing technique for tuberosity fixation were used in all operations. Patients were evaluated and radiographed with a minimum 2-year follow-up (mean 36 ± 8 months). The tuberosity union rate was 84% (32 of 38). There were 4 tuberosity resorptions and 2 tuberosity migrations with nonunion, which were associated with significantly lower subjective results (Subjective Shoulder Value of 65% vs. 83%, P = .029) and lower active mobility in forward elevation (115° ± 26° vs. 141° ± 25°, P = .023) and external rotation (11° ± 12° vs. 27° ± 12°, P = .010). Among the 5 disappointed patients, 3 presented with tuberosity resorption and 2 with tuberosity migration and nonunion. Despite the advanced age of the patients, tuberosity reattachment and use of bone graft results in a high rate of tuberosity healing. Tuberosity reconstruction and healing in reverse shoulder arthroplasty for fractures improves active forward elevation, external rotation, and patient satisfaction.

Background. The indications for reverse shoulder arthroplasty (RSA) have been widely expanded, but only a few studies report the long-term survival of these implants. Our objective was to report the long-term survivorship of a large series of RSAs implanted for different etiologies. Methods. A retrospective multicenter study including all the RSAs was performed in six shoulder-specialized centers with at least 2 years of follow up. We reviewed 1611 RSAs, operated between 1993 and 2010, including 497 cuff-tear arthropathies (CTA), 239 revision RSAs, 188 massive cuff tears (MCT), 185 fracture sequelae (FS), 183 failed previous cuff repairs (FCR), and 142 primary osteoarthritis (POA). The mean follow-up was 5.6 ± 3.9 years (range 2–20). Results. Overall, 266 RSAs (16.5%) had at least one complication leading to 64 reoperations (4.0%) and 110 revision surgeries (6.8%). The most frequent complications were infection (3.8%), instability (2.8%), and humerus-related complications (2.8%). At 10 years, the survival without revision surgery was 91.0% in primary RSAs and 80.9% in revision RSAs for failed arthroplasty (p < 0.001). In the primary RSA group, MCT and FCR led to 10-year survivals for over 95% but fracture sequelae and tumors had the lowest 10-year survivals (83.9% and 53.1%). Younger patients had a lower 10-year survival. In revision RSAs, male patients had a significantly lower survival than females (72.3% vs. 84.5% at 10 years, p = 0.020). Discussion. Primary RSA for cuff-deficient shoulders or POA leads to a high 10-year survival, but revision RSA or primary RSA for FS and tumors are at high-risk for revision. Surgeons should be aware of high rates of complications and lower survival rates of RSA in younger patients, in males, and in RSAs for revision surgery.

The high rates of complications and reoperations observed with the early designs of first-generation (unlocked) and second-generation (bent design) humeral intramedullary nail (IMNs) have discouraged their use by most surgeons. The purpose of this study was to report the results of a third-generation (straight, locking, low-profile, tuberosity-based fixation) IMN, inserted through a percutaneous approach, for the treatment of displaced 2-part surgical neck fractures. We performed a retrospective review of 41 patients who underwent placement of a third-generation IMN to treat a displaced 2-part surgical neck fracture (AO/OTA type 11A3). The mean age at surgery was 57 years (range, 17-84 years). After percutaneous insertion through the humeral head, the IMN was used as a reduction tool. Static locking fixation was achieved after axial fracture compression ("back-slap" hammering technique). Patients were reviewed and underwent radiography with a minimum of 1 year of follow-up; the mean follow-up period was 26 months (range, 12-53 months). Preoperatively, 3 types of surgical neck fractures were observed: with valgus head deformity (Type A = 8 cases), shaft translation without head deformity (Type B = 19 cases), or with varus head deformity (Type C = 14 cases). At final follow-up, all fractures went on to union, and the mean humeral neck-shaft angle was 132° ± 5°. We observed 2 malunions and 1 case of partial humeral head avascular necrosis. No cases underwent screw migration or intra-articular penetration. At last review, mean active forward elevation was 146° (range, 90°-180°) and mean external rotation was 50° (range, 20°-80°). The mean Constant-Murley score and Subjective Shoulder Value were 71 (range, 43-95) and 80% (range, 50%-100%), respectively. Antegrade insertion of a third-generation IMN through a percutaneous approach provides a high rate of fracture healing, excellent clinical outcome scores, and a low rate of complications. No morbidity related to the passage of the nail through the supraspinatus muscle and the cartilage was observed. The proposed A, B, and C classification allows choosing the optimal entry point for intramedullary nailing.