Ronald L. Linscheid, MD
In 1910 Kienbock' described the progressive dete-rioration of the lunate bone as seen by x-ray as luna-tomalacia and suggested that the etiology was injury to the blood supply of the bone. The next important ob-servation was that of Hulten who, in 1928, noted that in patients with Kienbbck's disease there was more likely to be a discrepancy in length between the ulna and radius than there was in the general population. This discrepancy was for the ulna to be shorter than the radius as measured at the dense cortical line of the lunate fossa perpendicular to the longitudinal axis of the forearm. This observation suggested the possibility of a mechanical susceptibility to this problem. It. also suggested a mechanical solution, for which he proposed radial shortening. Persson described an u1nar length-ening procedure in 1945, and others soon followed with reports of radial shortening procedures. The aims of both procedures were to redistribute the forces from the injured lunate towards the triquetrum. The vascular supply to the bone was also of obvious importance. In 1963, Lee proceeded beyond the ob-servation of the foraminal openings in the dorsal and palmar poles of the lunate to study the intraosseous vascular distribution within the bone. This suggested an increased susceptibility to avascular necrosis in those individuals with deficient anastomotic patterns. A num-ber of authors have made significant advancements in the study of the natural history of this problem. An-tuna-Zapico,s in a brilliant monograph in 1966, looked at the morphologic variations of the lunate and the stress concentration considerations this posed for the injured lunate. More recently, the carpal collapse as-pect of the problem has suggested additional treatment plans, but before proceeding further a look at the nat-ural history of the problem may be in order.
Natural History
The incidence of Kienbock's disease is greatest in the active young adult population but it can occur in child-hood and well into the later decades of life. It affects a higher proportion of men than of women. The initial symptom, aching in the wrist with activity, progresses to increasing limitation of motion, weakness, and sharp pain with movement. Some patients may tolerate these symptoms for years, and, occasionally, the symptoms slowly ameliorate to where function is not severely im-paired. It is partly for this reason-that the various treatments of this condition are difficult to compare objectively.
The pathologic changes that occur vary somewhat, depending on the etiology, the shape of the lunate, and the configuration of the radioulnar articulation. Sclerosis of the bone secondary to avascularity renders the bone susceptible to shear stress fractures in the subcortical trabeculae. These fractures usually occur parallel to the proximal articular surface and are similar to those seen in avascular necrosis of the hip and similar osteochondroses." Collapse of the trabeculae leads to compressive shortening of the lunate. If there are trans-verse fractures in the coronal plane, these same forces acting through the capitate tend to extrude the two fragments of the lunate dorsopalmarly. There is usually more involvement on the proximal convex aspect than on the distal concave surface, and flattening of the former is evident. The articular cartilage generally re-mains in good condition, but it is pliable because of the collapsed bone beneath. Collapse progresses until the joint compressive forces are attenuated by redis-tribution to the proximal scaphoid and triquetrum. The u1nar aspect of the lunate, which overlies the tri-angular fibrocartilage, is usually less involved in the collapse than is the portion that articulates with the lunate fossa of the radius. This variation is most readily explained by the difference in compliance between the two surfaces, especially in those with greater ulna minus variance. Shortening of the lunate may lead the scaph-oid to adapt a progressively more flexed attitude as it supports more of the joint compressive load, and this flexion is used to justify the method of treatment in which the scaphotrapeziotrapezoidal joint is ar-throdesed. For the most part, the scapholunate and lunotriquetral ligaments remain intact, helping to pro-vide transverse stability to the carpus. There are ex-ceptions to this, however, in which there appears to be concomitant scapholunate dissociation and lunatoma-lacia. Recent evidence suggests that the former con-dition is also more likely to occur with ulna minus. variance. Therefore, it is not surprising to see the simultaneous expression of both conditions.
Diagnosis
A history of injury is elicited in over 50% of patients. Even though the injury may have occurred months or even years earlier, there is reason to believe that the expression of osteonecrosis is often delayed, as will be explained later. The injury is usually a fall on the ex-tended hand.
The physical findings are limitation of motion, weakness, and point tenderness over the dorsal central as-pect of the wrist. Palmar flexion is more likely to be limited than dorsiflexion because of extrusion of the palmar pole of the lunate. Grip strength is often re-duced to 50% of the opposite hand.
Radiographic findings are based primarily on plain films of the wrist taken in the anteroposterior and sag ittal planes. The earliest finding is usually an increased radiodensity of the lunate as compared with the adja-cent bones. This increase is followed by a ra-diolucent line parallel to the articular surface through the subcortical area. Flattening of the articular surface, particularly over the lunate fossa, occurs with resorp-tion of trabecular bone, and the height of the lunate becomes noticeably reduced . Further fragmen-tation is followed by increasing signs of degenerative arthritis. This sequence of change is formalized in the staging plans of Stahl' DeCoulx and associates," and Lichtman and associates." Surprisingly little attention is paid to the sagittal views in most reports, which is unfortunate because these often provide important in-formation. The superimposition of the radial styloid, scaphoid, triquetrum, and ulnar styloid make interpre-tation difficult. Transverse fractures in the coronal plane, dorsopalmar extrusion, midcarpal angulation, and scaphoid angulation, however, may be seen with careful inspection.
Polyaxial tomograms are even mote informative, es-pecially when taken at intervals of 2 to 3 mm. I find that the stage of the disease is often upgraded when tomography and plain films are compared . Frac-tures are much more apparent, and so are the degree of flattening, the displacement of fragments, and carpal alignments. Although it is difficult to be certain unless an unequivocal fracture is seen in the lunate radio-graphs shortly after an injury, I believe that the coronal fractures are generally responsible for initiating the avascular changes. These coronal fractures may be sub-tle fractures at the palmar or dorsal poles or more obvious fractures through the body of the bone. The subcortical fractures in the sagittal or transverse planes appear to be more likely the result of shear fractures through avascular bone. Widening of the transverse fracture gaps by intrusion of the capitate is readily ap-preciated on these films, and it can also be seen on computed tomographic or magnetic resonance imaging scans. It is not unusual to see the palmar pole dis-placed so that it impinges against the palmar rim of the radius, which helps explain the loss of palmar flexion.
The stance of the scaphoid is also easier to appreciate on trispiral tomograms. In the neutral position of the wrist, the radioscaphoid angle is usually between 40 degrees and 60 degrees. Angulation much beyond this suggests some collapse of the carpus. This carpal collapse may be caused by collapse of the lunate, by midcarpal angulation, or, as is sometimes noted, by fragmentation of the palmar aspect of the lunate with proximal displacement and palmar subluxation of the capitate. To be sure, the scaphoid may also undergo angulatory displacement under dynamic loading, but unloaded static films suggest that only a small per-centage of scaphoids are abnormally flexed.
Because u1nar variance measurement is important in assessing the status of lunatomalacia and in its treat-ment, it should be done in a standard and reproducible manner. The wrist should be in a neutral position flat on the x-ray plate with the elbow flexed to 90 de-grees and the shoulder abducted 90 degrees. A line perpendicular to the longitudinal axis of the forearm is drawn from the proximal cortical line of the lunate fossa over the u1nar head. The distance to the cortical surface of the ulnar pole is measured. If avail-able, the concentric circle template recommended by Palmer is desirable, but the method must be consist-ent. Flattening or abnormal enchondral bony growth of the lunate fossa as a result of the lunate collapse has been noted minimally only in late cases.
Technetiurn 99 bone scans may provide an early in-dication of lunate injury, but localization may be im-precise and, if the scan is positive, it is best followed by tomography or magnetic resonance image scanning. The latter has the unique advantage of suggesting avas-cular changes in the bone long before they can be seen by radiograph. There is also the advantage of detecting articular effusion as well as fractures and the status of the adjacent carpus. This is likely to be an increasingly important tool both for diagnosis and followup, par-ticularly if such examinations become less expensive.
Etiology
The mechanism of injury to the vascular supply of the lunate is not well understood. The vessels can enter the bone only at the dorsal and palmar poles, because the bone is otherwise covered by articular cartilage ex-cept at the attachment of the interosseous membranes. The entering vessels execute a sharp bend as they come off the proximal carpal arcades. The studies of Lee and, more recently, of Gelberman and associates have shown three general categories of vascularity. These categories, described as Y, X, and I patterns, are marked by a decreasing intraosseous anastomotic network.
At both extreme extension and flexion the foraminal areas may impinge against the radial rims. There seems no doubt that interference with these vessels accounts for many cases of Kienb6ck's disease, because, in my experience, at least 40% of patients show no evidence of prior fracture. In this group of patients, sclerosis and proximal shear fracture or collapse of the convex surface are less likely to show extrusion lengthening ofthe lunate.
There is evidence of transverse lunate fractures oc-curring at the time of injury and progressing to fragmentation in a matter of several weeks. In other in-stances, however, a known fracture has not resulted in the secondary changes for several years or has persisted as a nonunion.
It is interesting to speculate why, when the lunate and scaphoid undergo essentially similar lesions, only a small percentage of scaphoid proximal poles undergo necrotic fragmentation bux a large percentage of frac tured lunates do. One answer may be that the stress concentration caused by the direct pressure of the cap-itate becomes much higher at the acute angle between the articular surface and fracture facet in the lunate than at the corresponding situation in the scaphoid. The more closed convex configuration of the cortical envelope of the scaphoid may also be much more re-sistant to compression.
Treatment
The ideal treatment is either to prevent deformity or to restore the lunate to normal appearance and func-tion. The latter has been an elusive goal. To accomplish this task in a stage III or IV situation would require restoring the height of the bone proximodistally, re-ducing the length dorsopalmarly, restoring the proxi-mal spherical convexity, re-establishing the vascular cir-culation, and preventing recollapse of the bone during the soft healing period. Some attempts to accomplish this using external distraction apparatuses, bone graft-ing of the trabecular defects, and introduction of a pedicled artery and its venae communicante have met with limited success. The healing period may readily extend beyond the wrist's tolerance for mechanical dis-traction, anatomic reduction is difficult, and ligated vessels may not maintain a flow capacity especially within the confines of trabecular bone graft fragments. If the initial susceptible condition is not corrected, the prob-lem can recur. Continued research in this area is, however, to be encouraged.
Prevention of Kienbock's disease can be accom-plished only when there is early suspicion after injury, with demonstration of a lunate fracture through the body of the bone. Unfortunately this opportunity is usually missed because the wrist is placed in a cast. Why this eminently sensible approach is likely to fail is ex-plained by the propensity of the capitate to force the fracture facets apart even with the wrist limited in motion. This problem is caused by the persistent joint com-pressive forces induced by the tension of the wrist and finger tendons still intermittently contracting within the cylindrical confines of a cast. If these forces can be removed from their impingement on the lunate, the stress risers at the sharply angled fracture edge will be reduced and the fracture facets are more likely to coapt.
The inability of the subcortical trabeculae to sustain the increased stress, especially when devascularized, can lead to collapse. The lunate fracture may heal even with a reduced vascularity and improve the late results. Even in that group in which no initial fracture exists, the earlier the lunate is unloaded, the less collapse is to be anticipated. For this reason, early decompression should be considered rather than conservative obser-vation in plaster.
Excision of the fragmented lunate, the oldest surgical treatment, has been effective in reducing pain at least initially. However, over time, the capitate intrudes into the open space and displaces the scaphoid and trique-trum, which leads to progressive degenerative arthri-tis To prevent this collapse, a variety of lunate re-placements have been used, including metal spheres, acrylic, silicone, and titanium models of prosthetic lun-ates. Biologic replacements have included capsular suspension, coiled tendons ("anchovies"), carved chon-dral cartilage, and pedicle pisiform bones. The dis-advantage of lunate excision is the disruption of the perilunate ligamentous supports, which favors displace-ment of the scaphoid, u1nar translation of the carpus, and instability of the prosthesis. Dislocation of the pros-thesis through the floor of the capsule is common un-less integrity is maintained by retention of a bony rem-nant of the palmar pole. Harder materials have a tendency to erode into the radius. Softer materials often deform or degrade to produce inflammation-ex-citing microfragments. All of the above treatments, however, have enjoyed a measure of success for varying periods.
So-called leveling procedures date from the early attempts of Persson to decompress the lunate by length-ening the ulna. Some patients treated by the early methods had follow-ups as long as 30 years with gen-erally satisfactory results. Better internal fixation de-vices, which allowed distraction lengthening with greater precision, used an inserted wafer of iliac crest or a grafted step-cut extension.
The obvious disadvantage of requiring a bone graft, along with the occasional nonunion, led others to adopt radial shortening as the preferred procedure. Both ul-nar lengthening and radial shortening are based on the premise that the u1nar head and the triangular fibro cartilage are able to distribute more of the compressive force through the triquetrum and u1nar aspect of the lunate .The latter was generally less affected than that part residing on the lunate fossa. Relative lengthening of the tendons also diminishes the overall joint compressive force. Another premise, later generally confirmed, was that the triangular fibrocartilage was thicker and usually unperforated in patients with an ulna minus variance. Thus, this fibrocartilage provides a compliant pad that supports the u1nar carpus. For-tunately, the distal radioulnar joint is quite forgiving to longitudinal displacement, except in those instances when it lies at a markedly oblique angle. Correction to an ulna 0 or + I variance is usually satisfactory. This technique may be used with a zero variant as well, but the correction should be small, or the patient is likely to complain of u1nar impingement symptoms.
Other methods of decompressing the lunate include scaphotrapeziotrapezoidal arthrodesis, capitohamate arthrodegis, and capitate recession. The first of these three methods depends on extending the scaphoid to its normal static position of about 45 degrees of radio-scaphoid obliquity so that the proximal pole accepts the primary joint compressive load. The sec-ond method assumes that the proximal migration of the capitate will be prevented by fusion to the hamate. The third method decreases the central loading through the wrist directly by shortening the capitate. All three methods have been afforded encouraging clin-ical reports.
Biomechanical studies suggest that the forces are re-distributed more evenly with forearm leveling proce-dures, and that scaphotrapeziotrapezoidal fusion is more effective than capitohamate fusion. Alter-ating the slope of the radial articular surface may also help levitate the lunate.
Salvage procedures, such as proximal row carpec-tomy, are occasionally warranted, especially when the distal surface of the lunate has collapsed or there is advanced collapse of the lunate as in the late stage III and IV." The contact area for the radiocapitate is considerably smaller than that of the radiocarpal because of the smaller radius of curvature; therefore, the compressive loading is high. The stability of the joint, however, is surprisingly good with an intact radiocapitate ligament. Isolated radiolunate fusion has not been effective, be-cause the avascular lunate does not heal to a graft.
Discussion
Long-term followups of the various methods of treat-ment are few and are not comparable in most instances. As noted, good results are reported by most techniques, a fact that may be explained in part by the tendency of the carpus to seek a state where further collapse has stopped and the wrist has stabilized. Little mention is made of the long-term status of the lunate or the car-pus. Healing with apparent revascularization may be seen in children and adolescents, but the proximal con-vexity of the bone does not reconstitute. In adults, free avascular fragments may be seen between healed por-tions of the lunate after as long as 17 years. The degree of original fragmentation and displacement may be a significant factor in this. Most treatment algorithms are based on the staging of Kienbock's disease as seen on the anteroposterior radiographs of the wrist. This cookbook type of ap-proach is better supplemented by a careful assessment of the multiple factors that may be involved. If polyaxial tomograms are not available, computed tomographic scans or a magnetic resonance image will give additional information. Study of the sagittal projections can often be more important than the anteroposterior view in assessing the lunate. Radial recession or u1nar lengthening to a 0 or + 2 mm variance, depending on the degree of initial variance, has the advantage of not de-stroying a normal carpal joint or interfering with in-tracarpal relationships. These treatments, which allow other methods to be employed later without compro-mise if the initial procedure is ineffective, may be com-bined with revascularization, bone grafting or fracture reduction. For this reason, this procedure appears to be the current benchmark against which other treat-ments are evaluated.