Why the elbow? My experience and perspective

Over the years I have frequently been asked why I decided to focus my attention on the elbow. My initial answer is that I really didn't in the sense that I continued to do hip and knee surgery throughout my career. But the specific reason for concentration on the elbow goes to my college years. I had a value-based education that emphasized service and the obligation to make the most of whatever talent you may have. I went to Mayo for my training because, for reasons that may not be obvious, I wanted to contribute to the specialty through basic and clinical research. The key year was 1972, the first year of my residency. When I was rotating with Dr. Patrick Kelly, Director of orthopaedic research, I asked what advice he could give me regarding a research project or direction. His cryptic answer was simply four words “work in a vacuum”.

There were two other events that directed me to study and treat the elbow. At the time of this discussion, I was doing a hand rotation with Drs. Linscheid and Dobyns. In a 6 week period we removed two failed silastic radial head prostheses. As I tried to read on the issue of radial head replacement, I discovered virtually nothing in the literature. In fact, the only reference to silastic radial head replacement was in a 1973 monograph by Dr. Al Swanson, inventor of the implant. The chapter dealing with the radial head referenced marketing material. When I reviewed the marketing material for its source of data, it referenced the monograph.¹ So not much help. As I attempted to learn more about radial head fractures, I learned there was very little written on the elbow at all, and only in one out of date textbook written by Dr. Fred M. Smith of Boston and published in 1972. Of interest there was no mention of elbow replacement.² This was not surprising as the elbow had been resistant to effective replacement, but implants were just beginning to be designed and the early results, reported in 1972, appeared promising.³ But already there were signs of early failure in almost all designs.

It was obvious that the early experience of partial or complete elbow replacement was without the benefit of a detailed biomechanical understanding of kinematics, force magnitude and articular transmission and relative contribution of articular and soft tissue constraints. I had found my vacuum!

1. My experience

With my NASA background as an aerospace engineer for over 2 years, and with Mayo laboratory facilities, I felt I might be in a position to address the problem. While in my second year of residency, under the tutelage of E Y Chao, PhD the recently appointed director of the biomechanics laboratory, I enrolled in the master's program of Biomechanics at the University of Minnesota. My master's theses project was a study of the kinematics of the elbow.⁴ Understanding elbow kinematics was a key step forward and for the first time we could appreciate the elbow motion as complex and unable to be replicated with a ridged hinge implant design. The study spawned the concept of a loose or ‘sloppy’ hinge articulation. This design feature was confirmed by a basic investigation showing comparable kinematic patterns of the normal and replaced elbow.⁵ The loose hinge is a feature of all currently used semiconstrained devices.

After serving 2 years in the Air Force in the wake of the Viet Nam experience, I returned to Mayo to practice as an adult reconstructive surgeon, which in those days included not just hip and knee, but also shoulder and elbow pathology. I immediately requested and received one day a week protected research time. It was my very good fortune to be able to work both with Dr. Chao, and also Dr Kai Nan An, PhD for the rest of my career. Since as mentioned we were in a vacuum, almost any area of basic investigation was the first of its kind and all findings represented some form of advancement of our understanding of the joint.

We structured our investigations to address the three categories of joint function: motion – kinematics; strength – force transmission and distribution; stability – soft tissue and articular constraints. We were fortunate to receive institutional and NIH support for our investigations- “Further study of kinematics of the unlinked devices and better understanding the stabilizing effect of the articular design”.⁶ These studies also allowed us to define (confirm) the location of the flexion axis and the instant center of rotation of the flexion arc. This insight led to the development of the dynamic external fixator. Subsequent studies allowed us to confirm the effectiveness of lateral half pin application to effectively resist both varus and valgus forces.⁷ Of practical value, the use of an electrogoniometer allowed the definition of the functional arc of the elbow to be 30–130° of flexion and 50° of protonation and supination. This has proven to be useful information when discussing the functional goals of surgery.⁸

Calculating the magnitude and direction of the forces across the elbow joint prompted the concept of an anterior flange applied to the humeral prosthetic component. It was theorized, and later proven that this simple concept could provide resistance to the greater than anticipated posteriorly directed forces during elbow flexion and extension. We subsequently calculated that this also markedly lessened the bone cement interface stresses that occur with torsion forces applied to the elbow.

A simple force/displacement experiment provided insight into the relative contribution of the ligamentous and articular surfaces to elbow stability under various loading conditions.^{9, 10, 11, 12, 13, 14}

These early experiences gave insight into the management of complex instability. A professional football player was referred to us because he was still unstable after undergoing the “Tommy John” medial collateral reconstruction. We had already identified a previously unrecognized instability mode whereby the ulna “rolled off” the humerus and the radial head slipped off the capitellum.¹⁵

Laboratory dissections had led us to identity a component of the lateral ligament complex that inserts onto the ulna. This component had gone unrecognized clinically as the band is covered by the insertion of the supinator on the ulnar crest. We had previously successfully reconstructed this structure and performed the reconstruction on an NFL football player. He returned to play the next season. It might now be recognized that this was in fact what we now call posterior lateral rotatory instability. When Dr. Shawn O'Driscoll did his fellowship, he had become interested in this condition and formally studied the instability pattern in the laboratory.¹⁶ He confirmed the role of the now termed lateral ulnar collateral ligament, to distinguish it from the “ulnar collateral ligament” which is the medial ulnohumeral stabilizer.

The common thread of all these inquiries was to provide a “scientific” basis for our clinical practice. Our basic investigations were designed to confirm, clarify, or define the proper management for the spectrum of elbow pathologies.

More recently, we have recognized the dominant, the genetically controlled “host variation” in determining the response to disease and to intervention. The laboratory has been investigating this phenomenon to identify the genetic controls of the process with the ultimate goal to develop a screen to identify those at risk for such conditions as arthrofibrosis and osteolysis due to nanoparticulate reaction.^17,18

2. Clinical investigations

Several overarching principles are followed when involved in clinical investigations. Studies are designed to use the volume and clinical retrieval competencies of the Mayo clinic with the goal to share our experience in managing the full spectrum of elbow pathologies. Hence, we have a requirement to publish our findings. We consider each joint pathology in one of three broad categories: acute trauma; soft tissue/sport (arthroscopy); and reconstruction for arthritis. In general I sought to address only issues that satisfied the following criteria: 1) important issue; 2) for which the answer was not known, or the accepted treatment may not be the best option; 3) Mayo had the volume and resources to answer the question definitively; 4) the project was ‘sized’ to the time available to study, hence projects taking more than a year to complete were assigned to residents, not fellows. 5) the project must be completed, and a draft of the manuscript finished before the resident or fellow leave Mayo.¹⁹ Finally, we tended to review our published outcomes periodically, e.g., every 10 years or so and update our experience as appropriate. I will attempt to briefly share what I think are some of the more important observations over the course of my career.

3. Trauma

We approach this subject with an effort to address the well-known reality that “the elbow does poorly”, you fill in the blank. We have attempted to provide simple and useful classification for the radial head fractures: simple or complicated²⁰; olecranon fracture: undisplaced, displaced, unstable – all with or without comminution; and the coronoid: Tip, <50%, >50%.²¹

Papandrea noted the unpredictable ability to salvage a terrible triad injury with ulnohumeral subluxation if the radial head fracture was not addressed.²² Concern exists regarding the second surgery required for radial neck fractures if a plate is used for fixation due to forearm rotation limitation. This led to the introduction of a ‘longitudinal’ fixation concept of placing screws at the margin of the radial head and engaging the cortex of the distal fragment. A controlled study revealed superior outcomes to plate fixation without the need for a second operation.²³

The two most impactful areas of involvement are that of understanding the terrible triad and its management, and the role of replacement for selected distal humeral fractures. One important insight is that trauma surgeons are typically not trained in elbow replacement, so some non-fixable fractures still undergo ORIF. The first publication of the expectation of replacement in the selected patient was that of Cobb et al., in 1997.²⁴ The short-term results were more than 90% successful. The experience was refreshed in 2017 reporting on 44 patients with a minimum of 10 year and up to 23-year surveillance. There were no humeral revisions for loosening in this period, but there were 5 ulnar revisions for loosening (3) or component fracture (2).²⁵

4. Sports, arthroscopy

Without question the emergence of elbow arthroscopy to address sports injuries is most impressive, but the continued development to assist in the fixation of articular fractures and address the common sequel of the stiff elbow is nothing short of spectacular. Much credit for this development goes to my Mayo friend and colleague Shawn O'Driscoll and my longtime friend and colleague Buddy Savoie.²⁶

Much of my experience with sports injuries involve instability and tendinopathy with cross over to the reconstructive realm with the management of the stiff elbow. Over time we have come to recognize that one source of stiffness, primary osteoarthritis of the elbow, is a “marginal disease”. That is to say, it is essentially a process of developing osteophytes at the margin of the articular cartilage. Hence its treatment is debridement, not replacement. The original approach to treat this condition was described as an ulnohumeral arthroplasty.²⁷ Although this provided very good, safe and predictable results, most primary arthritis with impingement is now equally effectively treated arthroscopically. The key to decision making and a good result, as noted by Antuna, is to recognize the potential for ulnar nerve symptoms and address them.²⁸ When stiffness does not involve the articulation and is not due to impingement, we developed the so called “column procedure” as a safe and effective means of predictably restoring motion.²⁹ If the stiffness involves the articular surface in the younger patients, I continue to employ the interposition arthroplasty and to date have not found an alternative to this, other than replacement which, as noted, is typically to be avoided.³⁰

Regarding elbow instability, we recognize that excluding sports, medial laxity is generally well tolerated, but lateral insufficiency leads to pain and dysfunction in activities of daily living and most of our attention has been directed to diagnosis and treatment of PLRI.³¹ One of the few contributions to the management of the medial ligament was to prove in the lab that, when removing the medial olecranon osteophyte of the thrower's elbow, removing as little as 3 mm of normal posterior-medial contour results in increased strain of the anterior bundle of the medial collateral ligament.³²

One additional experience that is not well recognized, is that if too much of the posterior medial corner is removed, repair of the anterior bundle of the ligament is not effective. In this setting I repaired the posterior bundle in one professional pitcher that had 2 failed anterior bundle surgeries. He returned to the Majors with a 3 year contract.³³

Poorly recognized early in my career, distal biceps injuries are now commonly diagnosed and well treated. Most of the literature has related to technique. My position is: do what works for you but be honest with yourself. I did a recent literature search and found that the functional results are comparable with the bone tunnel, anterior endo button and anchor procedures. However, the expense and the complication rates are statistically greater with the anterior approach. In my career, using the two incision, bone tunnel approach, I have not had a complication of heterotopic bone that caused any function limitation. The one additional useful insight recently documented by M Morrey is that in delayed cases, the repair can be performed in flexion of up to 90° and with time the tendon will stretch with excellent function.³⁴

On the other hand, rupture of the triceps is really quite uncommon. The diagnosis may be deceiving since the central tendon is almost always involved, uncommonly with osseous avulsion, and the patient can often extend against gravity. The best treatment is repair to bone tunnels assuming this repair is accomplished before contracture. The one major difference with the biceps is that triceps repair takes much longer to recover. I tell patients it may take up to a year to completely recover. For the well fixed biceps, I allow motion in a few days with gradual increase to full activity as tolerated by 3 months or sooner.³⁵

Finally, we have known for some time that the tennis elbow is a tendinopathy with apoptosis, not a cellular inflammatory process.³⁶ However, it does seem to be a chemical inflammatory process with the neurovascular controlled release of cytokines. My management of medial or lateral tendinopathy has changed in the last ten years to exclusively using the ultrasonic probe that aspirates the disease tendon and stimulates healing. The results have been about 90% successful in my practice without complication in treating almost 200 patients over the last 10 years.³⁷

5. Joint reconstruction

The strides made over the years are both impressive, and modest. Arthroscopic techniques to address intraarticular pathology, especially joint stiffness and cartilage repair is truly amazing. Other than the design features mentioned earlier, the greatest recent advance has been in the improved material characteristics of cross linked, vitamin E treated polyethylene. Unfortunately, there are few alternatives to replacement. Unlike other joints, there is no optimum position for fusion, just “least worse”. We have studied the proposition of shoulder compensation for limited elbow motion and discovered that ball and socket joint of shoulder cannot compensate for loss of hinge motion.^{38, 39, 40}

In my practice, the discovery of the value of the achilles tendon allograft has improved outcomes and certainly facilitated ligament repair. At a mean of 8 years, 88% said they would have the procedure again.⁴¹

As with any procedure, should the initial intervention fail the expectations of the salvage must be considered. We have reviewed the outcome of both a second interposition as well as implant salvage for a failed interposition. When the interposition was repeated, at an average of about 10 years, about 50% improved and about 50% underwent a replacement within 2 years.⁴² Fortunately, the replacement is not markedly affected by the prior procedure with about 90% satisfactory at 5 years, however, fewer excellent results were documented compared to the control group.⁴³

This brings us to our experience with joint replacement.^{44, 45, 46}

Having done over 1000 procedures and combining all Mayo Rochester colleagues experience we have documented our experience by presenting diagnosis. If the numbers allow, we have two studies on many of the diagnoses.

Rheumatoid arthritis - Gill performed a 12 year minimum study and revealed 91% were without revision at the time of the review.⁴⁷ The definitive study of almost 461 cases by Sanchez Sotelo demonstrated a sustained good outcome in these patients with 90% free of revision between 2 and 23 years.⁴⁸

Results of TER in acute fractures have been noted above. In established post traumatic arthrosis results have been reported in 1991 and 1997. The most recent review in 2010 by Throckmorton et al. revealed 84 patients followed for an average of 10 years with a satisfactory rate of 80%.^49,50 TER in post-traumatic nonunions have similar outcomes. Initially reported in 1995 with greater than 90% good results at 3 years, the results were shown to deteriorate with time. At a mean of almost 7 years, 93 patients had a subjective satisfactory rate of 85% but only 78% had satisfactory MEPS.^51,52

Revision TER - I think most of my personal growth in the last two decades is in the management of the osseous deficient elbow. Resection was reserved for the infected joint and we found it to be a poor salvage option with only 40% satisfactory outcomes at follow up.⁵³ The ultimate construct to address the problem of bone deficiency is to use an allograft prosthetic composite (APC). After an initial experience with marginal outcomes (55% success)⁵⁴ the system was modified. Using extended allograft struts, M Morrey reported more reliable fixation and the success rate improved to 85% at four years.⁵⁵ This remains a work in progress and it will remain so for some time to come. Hens it is the realm of revision that is most amenable to further development and improvement.