Epic Fail: Robots Don’t Improve Joint Replacement Outcomes in Big U.K. Studies

Robotically assisted knee and hip arthroplasties have brought no extra benefits for patients in Great Britain compared with conventional procedures, a massive review of registry data showed.

Revision rates over the first 5 years postoperatively were nearly identical for both total and unicompartmental knee replacement and for hip arthroplasties conducted with robotic aid versus standard manual surgery during 2018-2024, according to Hasan R. Mohammad, MRCS, DPhil, MRes, of Queen Mary University in London, and colleagues, who analyzed records for more than 2.5 million procedures conducted across the U.K., with the exception of Scotland.

Because the chief rationale for performing robotic arthroplasties is to improve implant positioning and thereby reduce malalignment and loosening that drive early revisions, it appears that robotic surgery may be a waste of money. In the group’s reports in The BMJ (one covering hip procedures and the other focusing on knee replacement), they estimated that each robotic device costs roughly £1 million to buy ($1.3 million at present) and another £1,000 to £2,500 ($1,300 to $3,300) for each procedure.

The studies may prove extra important in Britain as the country’s National Health Service (NHS) has announced plans to dramatically ramp up the use of robotic procedures. The U.K. government published a 10-year plan this past July under which the NHS would raise by sevenfold the proportion of hip and knee arthroplasties performed with robotics. “Beginning next year, we will expand surgical robot adoption in line with NICE [National Institute for Health and Care Excellence] guidelines,” the document stated.

NICE is a government agency that analyzes medical technologies for clinical efficacy and safety, and crucially, cost-effectiveness. Thus, if NICE accepts the new reports’ findings, it could well scuttle the shift to robotic procedures.

For the study, Mohammad and colleagues looked to the National Joint Registry of England, Wales, Northern Ireland, the Isle of Man, Guernsey, and Jersey. They identified the following numbers of procedures:

• Conventional total hip replacements (THRs): 656,080
• Robotic THRs: 10,203
• Conventional total knee replacements (TKRs): 585,242
• Conventional unicompartmental knee replacements (UKRs): 89,792
• Robotic TKRs: 17,431
• Robotic UKRs: 4,680

The researchers then propensity-matched each patient undergoing robotic procedures (not quite all of them, but almost) by clinical and demographic parameters, including surgeons’ procedure volume, to four who had conventional surgery, for a target trial emulation analysis.

In the matched THR cohorts, patients averaged 66 years old and 56% were women. Over 90% of surgeries were for osteoarthritis, with trauma and avascular necrosis accounting for most of the rest. Among TKR patients, the mean age was 70 and 56% were women; UKR patients were 66 on average and 54% were women. All but a handful of knee replacement patients were in for osteoarthritis.

Rates for 5-year implant survival (the inverse of revision) in the matched cohorts were as follows:

• Conventional THR: 98.8%
• Robotic THR: 98.8%
• Conventional TKR: 98.5%
• Robotic TKR: 98.6%
• Conventional UKR: 97.8%
• Robotic UKR: 96.4%

Findings were nearly identical when the entire unmatched populations were compared. For example, 98.6% of the THRs performed in the usual manual way survived, versus 98.8% with robotic surgery.

Mohammad and colleagues also examined all-cause mortality among THR patients, and found no significant differences by procedure type.

Robotic procedures did come with hints of advantages, which NICE may consider in developing guidance. THR revisions for dislocation were significantly less common in the unmatched analysis (0.11% vs 0.33%, P=0.002), as were acetabular and femoral fractures — but the differences shrank to insignificance in the propensity-matched cohorts. Knee replacements didn’t differ significantly in complications or specific reasons for revision, whether TKR or UKR, nor in the matched or unmatched groups.

It’s worth noting that revision rates aren’t the only measure of procedural success. Mohammad and colleagues observed that the better positioning and accuracy of robotically aided implantation may “lead to less pain and stiffness, better function, less joint wear, fewer fractures, and less instability.” However, these data suggested that such benefits weren’t big enough to reduce the need for revision.

“Given the observational design of the study … we cannot rule out the possibility of unmeasured and residual confounding,” the investigators acknowledged. Also, most procedures were performed in the NHS, which is notorious for long wait times and other factors peculiar to Britain that could limit the results’ generalizability to other countries and health systems.