Orthopedic Surgery

Midshaft clavicle fixation: better union, faster recovery, but no durable functional advantage

This systematic review pooled six RCTs with 579 midshaft clavicle fracture patients treated surgically or conservatively. Surgery produced higher union rates than nonoperative care and shorter time to union in most reporting trials. Short-term shoulder scores favored surgery, but there were no meaningful differences at or beyond 12 months. Overall complication rates were similar, with more hardware irritation after surgery and more nonunion after conservative treatment. These data support individualized, shared decision-making, emphasizing faster union versus similar long-term function and distinct complication profiles. 1

Preoperative hypoalbuminemia sharply elevates complications and mortality after THA/TKA

This meta-analysis included 14 observational studies and 1,194,088 total hip or knee arthroplasty patients. Preoperative hypoalbuminemia nearly tripled overall complications and quadrupled periprosthetic joint infection risk after arthroplasty. It also significantly increased sepsis, septic shock, pneumonia, myocardial infarction, wound complications, transfusion, and unplanned reoperation. Mortality risk was markedly higher in hypoalbuminemic patients, with an odds ratio of 7.14. The authors advocate routine albumin-based risk stratification and targeted preoperative optimization in THA/TKA candidates. 2

Supervised specific exercise clearly outperforms generic home exercise in subacromial impingement

This randomized trial assigned 94 adults with stage II subacromial impingement to 12 weeks of specific supervised exercise or general home exercise. Supervised therapy produced large Constant-Murley improvements over home exercise at 12 weeks and one year, exceeding minimal clinically important differences. Secondary outcomes (DASH, pain, kinesiophobia, irritability) similarly favored supervised exercise at both follow-ups. Responder rates and exploratory analyses consistently showed superior short- and long-term outcomes with supervised programs. These results support referring impingement patients for structured supervised rehabilitation rather than relying on unsupervised home regimens. 3

Deep learning system rivals senior surgeons for multi-abnormality hip radiograph screening

A ResNet-50-based model was trained on 25,908 hips and externally validated on 4,600 hips from another hospital. It classified eight categories including osteoarthritis, osteonecrosis, femoral neck and intertrochanteric fractures, dysplasia, and implants. External accuracy was about 90%, with macro-AUC 0.99 and strong sensitivity for femoral neck and intertrochanteric fractures. The model outperformed residents and attendings and was non-inferior to deputy chiefs for overall accuracy and macro-F1. Assistance improved surgeon accuracy by up to 11% and increased inter-rater agreement, supporting use as triage and second-reader support. 4