Humeral Head Replacement: When Half a Loaf Will Do

With the increasing popularity and availability of reverse shoulder arthroplasty systems for rotator cuff arthropathy, the indications for humeral head replacement (HHR) in such cases have decreased. Additionally, several studies have shown the superiority of total shoulder arthroplasty (TSR) over hemiarthroplasty for patients with glenohumeral arthrosis and an intact rotator cuff, especially with eccentric native glenoid wear.1 Finally, many severe proximal humeral fractures that, in the past, might have been treated with HHR are now being repaired with percutaneous fixation or newer locking plates. Thus, HHR is used less frequently. Current options in HHR include partial head replacement, humeral head resurfacing2 with or without fascial/meniscal interpositional allografts,3 and traditional stemmed hemiarthroplasties.4 Advances in design, such as partial replacement for focal defects and trabecular metal stems for improved tuberosity healing after fracture, may improve overall outcomes, allow for more individualized treatment options, and better define surgical indications in the future. What then, are the indications for HHR as opposed to TSR or, alternatively put, what are the indications and contraindications to implantation of a glenoid component? An easier way to frame the question may be to ask, “When is HHR the consensus; when is TSR the consensus; and when is TSR versus HHR controversial?”

Humeral head replacement has been the mainstay of treatment for severely displaced three- and four-part proximal humeral fractures since the advent of shoulder arthroplasty.5 Other indications for HHR include avascular necrosis with humeral head collapse and relative sparing of the glenoid; rotator cuff–deficient arthrosis in patients that are not candidates for a reverse arthroplasty6, 7 (since the reverse arthroplasty is a type of TSR); the physiologically young patient with glenohumeral degenerative joint disease in which the surgeon wishes to preserve glenoid bone stock for possible future TSR8; and some tumor reconstructions. In patients with glenohumeral arthrosis without sufficient bone stock for placement of a glenoid component, HHR is an option as well, and it should be noted that several of the above indications for HHR can be more accurately described as contraindications for glenoid component implantation.

Total shoulder replacement may be the consensus for elderly patients with advanced glenohumeral arthrosis affecting both sides of the joint, especially with eccentric glenoid wear.9 Even in this population, however, many surgeons are not inclined to resurface the glenoid due to the additional time required for glenoid implantation and the technical difficulty involved in glenoid exposure and preparation.10 In these cases there will always be debate, as these fall under the category of “surgeon preference” rather than strictly adhering to indications. Additionally, failed HHR is an indication for TSR, while cuff-tear arthropathy with pseudoparalysis and in intact deltoid is an indication for reverse TSR, which for purposes of this discussion will be considered a type of TSR, although some would consider it a unique entity.11, 12

More importantly, the question remains, “When is TSR controversial?” In younger, active patients with advanced glenohumeral arthrosis involving both sides of the joint, adequate bone stock, and an intact or reparable rotator cuff, TSR is still under debate and there are confounding trends that support both sides of the argument. Total shoulder replacements are getting better—especially with regard to the glenoid component. More anatomic designs and improved cementing techniques as well as improved glenoid preparation all contribute to improved results of TSR.13 At the same time, techniques of HHR also continue to improve. These include preferential reaming of uneven glenoids as well as alternative procedures such as interpositional arthroplasty3 or surface replacements,2 which are still options for younger patients.

A more appropriate question to ask is whether early results of TSR are better than early results of HHR and if there is a trade off or price to pay later (such as earlier loosening) for the improved early results of TSR. In fact, results and survival of TSR have been shown to be better than those of HHR in several retrospective as well as prospective studies.14, 15, 16 Additional excellent studies from Europe have shown better age-adjusted Constant scores, less pain, and better range of motion after TSR than after HHR.17, 18 Two prospective, randomized studies each showed better results with TSR compared with HHR,19, 20 and a metaanalysis of these two and a third not yet published showed that many HHRs later required conversion to TSRs.21

But, is later loosening the price paid for better early results? The recent literature does not support this. At a minimum 10-year follow-up and an average of 14 years, Deshmukh and coworkers22 from Boston showed 73% survival at 20 years. Sperling and coworkers’23 study showed better survival of TSR than HHR at 20 years in patients under the age of 50 (84 versus 74%, respectively) and many of the hemiarthroplasties required later conversion to TSR. These studies suggest that the problem of glenoid loosening and premature failure in younger patients may be less of a problem than previously believed.

For those less inclined to implant a glenoid component the alternative is for initial hemiarthroplasty and conversion to TSR down the road, if necessary. Unfortunately, the results of HHR conversion to TSR are less than encouraging. The study by Carroll and coworkers24 from Columbia showed only 20% excellent results in conversion of a previous HHR to a TSR and unsatisfactory results in 47% (7/15) of the patients. Additionally, in Sperling and Cofield’s25 study 39% (7/18) of patients had an unsatisfactory result, which makes the idea of initial HHR with the option of later conversion to TSR a less inviting option.

In conclusion, HHR is indicated for acute fractures and avascular necrosis without glenoid involvement. With glenoid involvement TSR has better early results than HHR for patients with an intact or reparable rotator cuff. Although durability in the long term may favor hemiarthroplasty, this difference may be less than previously believed and recent technical advances may improve results of both overall. So, we implant a glenoid component when the glenoid is diseased, the rotator cuff is intact or reparable (as in osteoarthritis or some rheumatoid patients), and the patient is willing to restrict activity and to comply with our rehabilitation protocols.