H. K. WATSON and P. M. GUEDERA
From the Connecticut Combined Hand Surgery Service, Hartford, Connecticut, USA
Journal of Hand Surgery (British and European Volume, 1997) 22B.- 1: 5-7
The aetiology of Kienbock's disease is unknown. The ideal treatment is in doubt. The purpose of this paper is to outline a hypothesis, drawing together much of what is known, for the aetiology of Kienbock's disease in the hope of stimulating further work and insight.
FAULT PLATE HYPOTHESIS
The factors influencing lunate necrosis are multiple, and a combination of them must be present for the condition to develop. These factors are related to unusual stress on the lunate over a period of time. They may be arbitrarily grouped as extrinsic factors, which are not features of the lunate itself but which act upon the lunate; or intrinsic anatomical features of the lunate which either render it unsuitable for load handling or which contribute to its failure under load. The result of a combination of these factors is the formation of multiple plates or faults within the substance of the lunate. These fault plates are comparable to geological faults and form in response to elastic deformation occurring in the trabeculae secondary to loading. The plates thus formed may be very small or very large. They may be localized to one area of the Innate or diffusely distributed throughout the bone. They may be short-lived or semi-permanent. Enough of them in crucial areas forming more rapidly than repair can occur can eventually wall off and interfere with blood supply to areas within the lunate resulting in bone necrosis . The anatomy of blood vessels per se probably plays little or no role.
EXTRINSIC FACTORS
Capitate
The radius of curvature of the proximal lunate on the radius is large and the bone is broadly and well supported. However, the radius of curvature of the capitate in the distal lunate is quite small. This concen-trates loads coming proximally through the capitate into a relatively small area and loads here tend to spread the lunate (Armistead et al, 1982). In a lunate with weakened internal structure secondary to bone cell death, the lunate will fail, usually producing palmar and dorsal segments. This fracture is probably a terminal event (Beckenbaugh et al, 1980), not an aetiological one. Lunate fractures are the result of Kienbock's disease, not the cause. The capitate, however, is the chief pile-driver in the engine of destruction resulting in Kienbock's disease.
Lunate loading
The tendency for the proximal pole of the scaphoid to displace dorsally under significant inline loading probably protects the scaphoid. The lunate, however, cannot escape the small head of the capitate and the loads are concentrated through the lunate even in extreme wrist flexion or extension. The lunate acts as a keystone (Marek, 1957), prevented from displacement through its attachments to the scaphoid and triquetrurn (Armistead et al, 1982). Proximally directed forces from the hand tend to center in the lunate, much as the keystone of an arch takes load at the apex of convexity. Avascular necrosis of the scaphoid may represent multifactorial overload and fault plate formation in a scaphoid identical to Kienbock's disease but is less common because the scaphoid can more easily share loads with its ligaments. In a sense, this is the scaphoid's salvation but clearly also is its downfall in terms of instability problems.
Ulnar variance
This is one of the most commonly cited factors in lunate loading. If the ulna is short, it seems reasonable to assume that it cannot share loading with the radius, thereby increasing the loads in the radial half of the lunate. The lunate may be specifically overloaded by the ulnar radial articular ridge, which acts as a fulcrum localizing loads sufficiently to produce fault plates. There are numerous studies to support this mechanism (Axelsson, 1971; Chen and Shih, 1990; Gelberman et al, 1980; Hulten, 1928; Stahl and Reis, 1986). Additionally, there are reports of Kienbock's disease associated with ulna positive variance (Nakamura et al, 1991). Here, the mechanism may be an impaction phenomenon from the ulna; either results in formation of fault plates. Interestingly, however, the lunate seldom fractures into radial and uInar segments over this ulnar-most edge of the radius. After necrosis, the capitate is responsible for fracturing the lunate into palmar and dorsal pieces. This means that fault plates may be produced by changes in the lunate-ulna relationship, but loads are not sufficiently concentrated by lack of uInar support to fracture the lunate along the uInar edge of the radius. This gross fracture of the lunate would also be a terminal event separate from the overloads within the lunate which produce shear injuries to the trabeculae resulting in the fault Plates.
Load type
The loads applied are aetiological (Joji et al, 1993). Sudden, high-impact loads such as jack-hammering seem obvious, but any heavy hand activity will suffice. Non-impact, high torque loads are magnified in the carpals by wrist flexion and extension. A mentally deficient patient who rocks incessantly on his hands falls into this factor group as much as the rough carpenter. There is also a possible relationship between Kienb6ck's and the use of vibratory tools (Gemme and Saraste, 1987; Letz et al, 1992). This may be caused by a direct vascular insult, but more probably is due to vibratory fatigue fractures of trabeculae resulting in fault plates.
Instability
Scapholunate dissociation, with subsequent destabilization of the lunate, has been reported to be associated with Kienbock's disease (Bourne et al, 1991; Cope, 1984). It is possible that instability is an extrinsic factor. The rotating scaphoid easily escapes to the dorsum. of the capitate, leaving the lunate to take the load. The lunate then remains loaded even at maximal wrist displacement positions (Armistead et al, 1982; Kashiwagi et al, 1977). The rotation of the scaphoid typically associated with Kienbock's is secondary to the lunate collapse and probably not causative, however (Watson et al, 1985; 1993).
INTRUNSIC ANATONUCAL FACTORS
Lunate spherical shape
The lunate is more spherical than most of the other loaded carpal bones and thus more dependent on trabecular than cortical support. For the same cortical thickness in similar sized bones, the more spherical the bone the more it will depend on trabecular support to prevent collapse under load.
Cortical strength
The lunate is physiologically ballottable. If one could actually measure the change in the shape of the lunate under heavy load, one would see a bulging particularly of the palmar and dorsal cortices and a compression of the distal cortex toward the proximal cortex. This compressive tendency of the lunate dictates a degree of deformation within its substance. The elasticity and strength of the cortical walls will determine the "collapsibility" of any given lunate under any given load (Koebke et al, 1989). Clearly, if "collapsibility" exceeds certain limits, trabecular fractures will occur along shear planes within the lunate.
Type "V" and "D" Lunates
Kauer's work notwithstanding (Kauer, 1980), there are significant numbers (231%) of lunates that are thinner palmarly than dorsally (type V as opposed to classic type D) when viewed on lateral projection. The wedge shape may play a role by preventing either dorsal or palmar lunate displacement (escape) at maximum dorsi-flexion load position of the lunate. Either V or D type lunates may be aetiological factors in the creation of fault plates.
Anatomy of the trabeculae
There are X-ray variations in lunate trabecular anatomy between individuals (Antuna, 1966). Some X-rays demonstrate coarse, heavy, well spaced trabeculae while others demonstrate a fine-grain homogeneous trabecular appearance. One or the other may be more prone to elastic deformation and fault plate formation. A simple comparison of the trabecular type in normal versus wrists with Kienbock's disease might be worthwhile.
Position and type of trabeculae
The ossification centre position and architectural alignment of bone within the lunate may predispose to a decreased load tolerance. The quality and morphology of osteoid and mineralization may play a role in fault plate susceptibility.
Vascular anatomy of the lunate
This has often been cited as the cause of Kienbock's disease (Gelberman et al, 1980; Marek, 1957; Williams and Gelberman1993). This seems obvious, but one has to make the transition from blood supply to bone death in a manner which fits the clinical appearance of Kienbock's disease. The only theory which seems to allow for all known components is the multifactorial load creation of fault plates which eventually wall off and interfere with capillary level blood supply (Goldsmith, 1925). It may be that the relative paucity of blood supply is a factor, as in the lunate with a single nutrient vessel, but how this produces Kienbock's disease remains obscure. What may be important, however, is the plane in which the vessels lie. If, for instance, the major vessels lie in a palmar to dorsal axis, the formation of sagittal fault plates would tend to devascularize across these planes, resulting in necrosis (Williams and Gelberman, 1993). There may be a particular vascular anatomy which plays a major role in aetiology but it is our impression that this is not the case.
ONSET OF KIENBOCK'S DISEASE
Kienbock's disease presents in various forms. The changes in the lunate may be very localized or diffuse. With progression, the lunate may fracture into two pieces or crumble into granules or produce an osteochondral fracture from one surface. The cartilage is not primarily involved in the pathogenesis and remains relatively healthy. Bone death occurs over time. Often, a symptomatic wrist will demonstrate normal X-rays with the diagnosis evident only on MRI (Desser et al, 1990). The degree of loading of the wrist appears to be important and changes in the lunate are cumulative.
A typical scenario is a lunate with several of the intrinsic anatomical features noted above, predisposing the lunate to elastic deformation of trabeculae. Microintratrabecular fractures occur and heal, but have certain changes along the plane of injury. With repeated injuries along the same plane and resultant rupture of capillary systems, there develops a physiological fault plate with increased resistance to normal blood flow (Jensen, 1993).
Additional fault plates may form in other areas of the lunate. This phenomenon is probably present in most of our carpal bones, but time and the avoidance of similar overload injuries preclude development of multiple plates. Areas in the lunate reach a critical state in which either a small section or much of the bone becomes relatively walled off by multiple fault plates and further "normal" injuries can no longer heal effectively because the capillary blood supply is inadequate. portion of bone then dies, providing a significant area further limiting vascular access.
Changing one or more of the actiological factors can halt progression and the necrosed areas walled off by fault plates can probably heal. Collapsed areas, however, remained deformed.
Most cases of Kienbock’s disease devekop, then, as the result of a long process of insult which is multifactorial in aetiology, producing overload within the substance of the bone. In the susceptible lunate, miniplanes of injury result in relatively avascular fault plates, gradually sectioning off areas in the lunate. Healing in these small areas cannot occur quickly enough if the abnormal loads and demand continue, ultimeately resulting in areas of cell death.