This Week in Orthopedics — Oct 1, 2026

This AudioScholar briefing is generated by artificial intelligence for healthcare professionals and trainees. It is not medical advice.

Welcome to This Week in Orthopedics. This week we're covering 10 notable papers spanning arthroplasty technology and alignment, knee preservation and rehabilitation, and the systems-level questions of cost, comorbidity and risk stratification. Let's dive in.

We start with robotics, where two large registry analyses published in the BMJ land on the same cautious note. Mohammad and colleagues used a target trial emulation design applied to National Joint Registry data, which is about as close as observational work gets to a randomised comparison. In the knee analysis, nearly 700,000 knee replacements performed between 2018 and 2024 were examined, of which about 22,000 were robotic [1]. With a mean follow-up of two and a half years, five year implant survival sat at roughly 98.5 percent in both the conventional and robotic total knee groups, and there was no difference in overall revision risk, cause-specific revision, or intraoperative complications. The same held for unicompartmental replacement. The hip paper covered about 666,000 total hip replacements, around 10,000 of them robotic, and again found no difference in overall implant survival, patient survival, or complications [2]. There was one signal: revisions attributed specifically to implant malposition, meaning dislocation, leg length discrepancy and malalignment, were roughly halved in the robotic group. That is a small absolute number of events and a secondary outcome, and the authors are explicit that residual confounding cannot be excluded. Taken together, these two analyses suggest that at early follow-up robotics has not yet translated into better implant survival, and the authors frame this as an argument for careful evaluation of a technology that carries substantially higher capital and procedural costs. Longer follow-up may change the picture, since revision curves for alignment-related failure diverge late.

If the robot has not yet proven itself, the question of how we align the knee remains live. In Knee Surgery, Sports Traumatology, Arthroscopy, Campi and colleagues report a retrospective comparison of 622 primary total knee arthroplasties, split roughly evenly between unrestricted kinematic alignment and mechanical alignment, all using the same implant and standardised perioperative protocols [4]. At a minimum of one year, overall satisfaction was 94 percent with kinematic alignment versus 81 percent with mechanical alignment, residual pain was roughly halved, and Forgotten Joint Scores were about nine points higher in the kinematic group. Stair climbing, swelling and stiffness did not differ. This is level three retrospective evidence from a single group, so selection and surgeon effects cannot be ruled out, but it adds to the case that alignment philosophy, rather than the tool used to execute it, may be what drives the so-called unhappy knee. Read alongside the registry data, the two findings sit together coherently: how you align may matter more for symptoms than whether a robot helped you get there, though neither paper tests that directly.

And the volume question is not going away. In the Journal of Bone and Joint Surgery, Adolf and colleagues applied a uniform Poisson modelling approach to registry data from 17 countries to project revision arthroplasty demand through 2060 [5]. Aggregate annual revision volumes are projected to rise by roughly a third by 2060, with revision knee growth outpacing revision hip. The trajectories are strikingly heterogeneous: the United States and China drive most of the absolute growth, while Germany, the United Kingdom and Sweden approach plateaus by mid-century and Japan is projected to decline. These are modelled projections, level four prognostic evidence, sensitive to assumptions, and the authors bracket that uncertainty with sensitivity analyses. Their point is about infrastructure: revision fellowship pipelines, infection prevention programmes, and registries themselves.

Turning to the younger knee, Annals of Internal Medicine published a randomised trial from Culvenor and colleagues addressing a genuinely under-served group: young adults with persistent knee problems nine to thirty-six months after anterior cruciate ligament reconstruction [3]. One hundred and eighty-four participants, mean age thirty, were randomised either to four months of twice-weekly physiotherapist-supervised exercise therapy with education, or to a single education session plus a self-directed exercise resource. At four months the supervised group improved more on the composite knee outcome score, by just under five points, with greater strength gains and considerably higher rates of patient-perceived success, around three quarters reporting improvement in pain versus about four in ten in the control arm. The important nuance is what happened next: by twelve months the between-group difference had narrowed, because the control group kept improving while the supervised group held its gains. There were no differences in cartilage thickness or composition on magnetic resonance imaging, so this is a symptom and function story, not a structural disease modification story. The trial supports supervised rehabilitation for faster symptomatic recovery in this population, while leaving open whether the added resource buys anything durable at a year.

Staying with knee preservation, Arthroscopy published a modified Delphi consensus from Super, Tollefson, Murray and colleagues, with 29 knee preservation specialists across five continents reaching agreement on 42 statements about unloader bracing for tibiofemoral osteoarthritis [6]. The panel supported bracing within a multimodal pathway for symptomatic unicompartmental disease, particularly where malalignment is correctable on examination, and endorsed its use as a screening tool before osteotomy, for symptom control while awaiting surgery, and as a postoperative adjunct. They agreed on a minimum six-week diagnostic trial and a six-month therapeutic period. This is level five expert opinion, and the panel themselves flagged standardised protocols and high-quality randomised trials as the priority, so this is a framework for practice variation rather than evidence that bracing changes outcomes.

Two papers in Spine turn to who gets worse outcomes and where. Kishan and colleagues analysed nearly 400,000 patients undergoing single-level degenerative spinal fusion in a national claims database, testing 28 preoperative comorbidities against revision within two years [8]. Anaemia, osteoporosis, substance-related disorders and depression carried the largest risks of all-cause revision, each raising it by roughly a fifth to a third, and those four were consistently associated with pseudarthrosis, mechanical failure and adjacent segment disease. Hypertension and type 2 diabetes conferred more modest increases, and notably were associated with mechanical failure specifically. Adjacent segment disease was uniquely linked to osteoarthritis. Strikingly, no comorbidity reached significance for infection-related revision. This is claims data with the usual coding limitations, but it supports comorbidity-informed risk stratification and points at modifiable targets like anaemia and bone density before elective fusion.

The second Spine paper, from McDevitt and colleagues, asks whether paying more buys better care [7]. Across nearly 242,000 elective anterior cervical discectomy and fusion encounters, hospitals in the highest regional decile of price markup, meaning billed charges relative to actual cost, had a mean markup of nearly eight times cost versus about four times elsewhere. After adjustment, those high-markup hospitals had modestly but significantly higher 30-day readmission, perioperative complications, long-term complications and reoperation. The absolute differences are small, a few tenths of a percentage point in each case, and this is retrospective administrative data, but the direction is clear: higher charges did not track with better outcomes in cervical spine surgery.

Two final papers address recognising harm and recognising risk on imaging. In the American Journal of Sports Medicine, Hoit and colleagues prospectively screened 93 patients after hip arthroscopy performed supine with a perineal post, using sensors to record traction force and duration and confidential questionnaires out to a year [9]. The incidence is the headline: about 83 percent of patients had symptoms consistent with pudendal nerve palsy in the acute period, most commonly perineal numbness, and urogenital or sexual dysfunction occurred in about 39 percent, including erectile dysfunction in roughly a fifth of men. At three months about a quarter of patients still had symptoms, 14 percent persisted past six months, and all had resolved by twelve. Each additional ten minutes of traction raised the odds of palsy by roughly sixty percent, and symptoms persisting beyond six weeks tracked with meaningfully worse hip-specific scores at six months. It is a single-centre cohort of modest size, but it suggests this complication is far more common than reported, and that traction time and force are the measurable levers.

And in the Journal of Bone and Joint Surgery, Bigness and colleagues externally validated Bone-RADS, the American College of Radiology framework for risk-stratifying lytic bone lesions, across 423 extremity radiographs scored independently by four reviewers [10]. Each one-point increase in score was associated with roughly a tenfold increase in the odds of malignancy. Sensitivity was about 97 percent with a negative predictive value above 90 percent, but specificity was only about 67 percent, and inter-rater agreement on the four-tier score was moderate, improving when collapsed to a simple benign-versus-malignant split. The authors position it as a practical triage tool, good at not missing malignancy, less good at avoiding unnecessary workup.

If you only have time for one paper this week, make it the BMJ knee robotics analysis [1]. It is the largest and most methodologically careful early look at whether robotic assistance improves what we actually revise knees for, and its null result reopens the value question for a technology many units have already bought.

Here is what this week's evidence adds up to in Orthopedics. Robotic hip and knee replacement has not yet demonstrated better implant survival or fewer complications at early follow-up in national registry data, with the single exception of malposition-related hip revisions, and longer follow-up is needed before the question is settled. Alignment philosophy in total knee arthroplasty continues to generate stronger signals for patient-reported symptoms than the delivery technology does, though that evidence remains retrospective. For young adults with persistent symptoms after anterior cruciate ligament reconstruction, supervised rehabilitation accelerates recovery over self-directed care but the advantage appears to fade by a year, and neither approach altered cartilage on imaging. Across spine surgery, modifiable comorbidities such as anaemia and osteoporosis predict non-infectious revision, and higher hospital charges did not correspond to better outcomes. Finally, pudendal nerve symptoms after post-assisted hip arthroscopy appear considerably more common than the literature suggests, with traction time as the strongest measured predictor.

That's your roundup for This Week in Orthopedics. The full transcript and references are available on the episode page in your AudioScholar library. This is an AI-curated summary — for clinical decisions, always consult primary sources and current guidelines. See you next week.

If this weekly briefing is useful, follow the show in your podcast app so new episodes arrive automatically. And if it saves you time, a quick rating in your podcast app helps other physicians find the show.

This is an automated summary generated by artificial intelligence, which can make mistakes. Always review the original source materials.