Proximal Scaphoid Reconstruction Techniques Comparison of Three Techniques for Proximal Scaphoid Recontruction

Abstract

Introduction  The costo-osteochondral autograft, vascularized medial femoral trochlear osteochondral autograft, and proximal hamate autograft have been used for the reconstruction of unsalvageable proximal pole scaphoid nonunions. Our hypothesis is that there is no difference in carpal kinematics after the proximal pole of the scaphoid is reconstructed with these three graft options.

Methods  Wireless sensors were mounted to the carpus that was loaded through cyclical motion. Each specimen was tested under a series of the three reconstructed conditions and their kinematics compared.

Results  No significant differences were found in scapholunate and lunocapitate joint motion during wrist flexion-extension and wrist radioulnar deviation between the three reconstructed conditions ( p  > 0.05).

Discussion and Conclusion  There are minimal differences in carpal kinematics when comparing reconstruction of the proximal pole of the scaphoid with the costoosteochondral, medial femoral trochlear, and proximal hamate grafts.

Scaphoid fractures account for approximately 60% of carpal fractures ¹ with a reported incidence of 1.5 to 39 per 100,000 person-years. ^{2 3 4 5 6 7 8} Gelberman and Menon demonstrated that the majority of the blood supply to the scaphoid arises from the dorsal ridge. ⁹ Additional vessels enter the distal tuberosity. This retrograde blood flow predisposes the proximal pole to vascular insufficiency and associated nonunion or avascular necrosis in the fracture setting. Fragmentation of the proximal pole in these scaphoid fracture nonunions poses a particularly challenging treatment dilemma. Over the last decade, there has been controversy regarding the optimal management of unsalvageable proximal pole scaphoid nonunions. ¹⁰ The costo-osteochondral graft, vascularized medial femoral trochlear (MFT) osteochondral graft, and proximal hamate graft have been used to address unsalvageable proximal pole scaphoid nonunions. Our hypothesis is that the proximal hamate graft equally restores carpal kinematics when compared to reconstruction with the costo-osteochondral and MFT grafts.

Methods

A sample of convenience of eight fresh-frozen cadaveric specimens were chosen for this institutionally approved study (Project ID: 18-006280). Specimens with wrist arthritis, previous wrist surgery, and type 2 lunates were excluded from this study. Pins and suture anchors were used to mount sensors to the cadaveric carpus. ¹¹ These sensors were positioned in the distal radius, lunate, capitate, scaphoid, and third metacarpal under fluoroscopic guidance ( Fig. 1 ). The flexor carpi ulnaris (FCU), flexor carpi radialis (FCR), extensor carpi ulnaris (ECU), and extensor carpi radialis longus (ECRL) and brevis (ECRB) were loaded during testing. ¹¹ As has previously been described, a wrist simulator moved the wrist through a cyclical motion about the flexion–extension and radial–ulnar deviation axes. ¹¹ The specimen was mounted to an unconstrained X-Y table using a clamp on the hand, while the distal forearm was mounted with Kirschner-wires (K-wires) passing through the radius and ulna to a motor-driven stage that moves to create the desired wrist motion ( Fig. 2 ). A joint compressive force (15 N) was applied using four pneumatic actuators applying tensile forces to the five tendons: FCU, FCR, ECU, and ECRL/B.

Fig. 1.

Cadaveric hand mounted to wrist simulator, ready for movement about the radial–ulnar deviation axis with wireless tracking markers mounted to the third metacarpal, radius, scaphoid, lunate, and capitate. Hand fixed to free-sliding X-Y state, forearm gantry mounted on motor-driven rotary stage, pneumatic cylinders apply constant, passive load across wrist.

Fig. 2.

( A–D ) Scapholunate motion angles during wrist range of motion in flexion–extension and radial–ulnar deviation (Hamate: proximal pole of the hamate graft, knee: medial femoral trochlear graft, rib: costo-osteochondral graft).

Each specimen was tested under a series of three conditions: (1) reconstruction of the proximal pole of the scaphoid using a proximal hamate graft, (2) reconstruction of the proximal pole of the scaphoid using a MFT graft, and (3) reconstruction of the proximal pole of the scaphoid using a costo-osteochondral graft. ^{11 12 13 14} An intact wrist condition was not tested. Within each condition, the specimen was evaluated via cyclical testing about two functional axes of motion (flexion–extension and radial–ulnar deviation.) The wrist underwent 100 cycles for each axis of motion at 70 degrees/s. The position and orientation of the hand (tracked via a sensor on third metacarpal), scaphoid, lunate, capitate, and forearm (tracked via a sensor on radius) were recorded at 100 Hz using motion capture software (The Motion Monitor, Innovative Sports Training, Inc., Chicago, IL) for the final five cycles of motion.

Kinematic motion was captured with the Moiré Phase Tracking three-dimensional (3D) motion tracking sensor hardware (MPT, Metria Innovation, Inc., Milwaukee, WI) and motion capture software (The Motion Monitor, Innovative Sports Training, Inc., Chicago, IL) to evaluate the hand, wrist, and forearm kinematics. This system utilizes a single camera and a single, passive, wire-free marker containing a unique Moiré pattern to track the position and orientation of each rigid body in 3D space. Data were collected at 100 Hz using The Motion Monitor toolbox software (Innovative Sports Training, Inc., Chicago, IL). The position and orientation accuracy of the system are 0.4 mm and 0.05 degrees, respectively. These, small, sensors enabled accurate recording of the carpal motion without having a confounding effect on the carpal motion itself. The anatomical coordinate systems of the hand and forearm were defined using a calibrated digitizing stylus according to the International Society of Biomechanics standards. ¹⁴ The coordinate systems of the capitate, lunate, and scaphoid were aligned to that of the hand (tracked via third metacarpal sensor) with the wrist in the neutral position. The orientation of these directly tracked coordinate systems was then used to compute joint angles via Euler angles. Euler angles were computed with the rotation sequence about the mediolateral (flexion–extension), anterior–posterior (radial–ulnar deviation), and superior–inferior (pronation/supination) axes. All coordinate systems were defined with the wrist in its native state prior to any surgical interventions and the tracking sensors remained on throughout the entire data collections and all surgical interventions. Therefore, the measured carpal kinematics remain absolute between conditions (i.e., are not reset to a relative neutral after each surgical intervention).

The last 5 of the 100 cycles performed were averaged and used for kinematic analysis. Scapholunate (SL) and lunocapitate kinematics were evaluated during wrist flexion–extension and radial–ulnar deviation motions for each of the three conditions. The lunocapitate and SL kinematics were fit to discrete, integer values of wrist flexion–extension and radial–ulnar deviation angles using piecewise cubic interpolation to enable statistical analyses to be performed at each discrete wrist angles. Differences between wrist conditions at each 5 degrees interval were used for the statistical analysis.

Surgical Technique

We made a longitudinal dorsal skin incision and approached the carpus through the extensor retinaculum between the third and fourth extensor compartments. A ligament sparing capsulotomy was then performed preserving the dorsal inter carpal ligament. Under fluoroscopic guidance, we used an osteotome to create a fracture of the proximal pole of the scaphoid.

The proximal pole of the scaphoid was sequentially reconstructed in a randomized order with a proximal pole of the hamate graft, MFT graft, and costo-osteochondral graft. The graft was harvested following the original description of each technique. ^{11 15} Once satisfactory reduction was confirmed with direct visual inspection and with fluoroscopy, the graft was fixed with a 2.5 mm headless compression screw (Arthrex, Naples, FL). Fluoroscopy was used to confirm appropriate reduction and fixation of the graft using multiplanar views. When using the proximal pole of the hamate graft, we repaired the SL ligament by suturing the volar capitohamate ligament to the lunate part of the remaining SL ligament as described for the hamate and for the rib graft, the dorsal SL ligament was sutured to the costochondral cartilage. ¹¹ The dorsal capsulotomy including the dorsal intercarpal and radiocarpal ligaments and the extensor retinaculum were repaired and the specimen was mounted in preparation for testing after each reconstruction.

Statistical Analysis

A multivariate repeated measures analysis of variance was performed with p  = 0.05 comparing the SL and lunocapitate angles at each 5 degrees interval of wrist flexion–extension and wrist radial–ulnar deviation between the three carpal conditions (reconstruction with proximal hamate, reconstruction with MFT graft, and reconstruction with costo-osteochondral graft) with appropriate Greenhouse-Geisser corrections applied based on the results of Mauchley's sphericity tests. A Tukey honestly significant difference post-hoc test was performed in the event of statistical significance.

Results

No significant differences were found in SL motion during wrist flexion–extension and wrist radioulnar deviation between the three reconstructed conditions ( Fig. 2 ; p  = 0.05–0.007). There were also no significant differences in lunocapitate flexion–extension and radioulnar motion during wrist flexion–extension and wrist radioulnar deviation ( Fig. 3 ; p  = 0.05–0.007). The intact condition from a predicate study is included in Figs. 2 to 3 for visual comparison but accurate statistical comparisons could not be performed due to the repeated measures approach and different specimens were used between the studies. ¹⁴

Fig. 3.

( A–D ) Lunocapitate motion angles during wrist range of motion in flexion–extension and radial–ulnar deviation (hamate: proximal pole of the hamate graft, knee: medial femoral trochlear graft, rib: costo-osteochondral graft).

Discussion

The purpose of this study was to compare carpal kinematics after reconstruction of the proximal pole of the scaphoid with either a proximal hamate, MFT, or costo-osteochondral grafts. We hypothesized that there would be no significant difference in carpal kinematics between the three reconstructed conditions. Overall, we found minimal differences between the three reconstructed conditions. This is important for the surgeons when considering which graft to choose from when considering either of these procedures.

Several vascularized and nonvascularized autograft and allograft options for reconstruction of the scaphoid proximal pole have been described with varying degrees of success. ^{12 13 14 16} Carter et al reported their experience of reconstructing the proximal pole with a scaphoid allograft. ¹⁷ Their technique required transforming the proximal pole fracture into a waist fracture to allow for reliable fixation of the allograft to the native scaphoid with a headless compression screw. In their preliminary report, the authors reported satisfactory healing, pain relief, and range of motion in six of the eight cases. Sandow described a technique for proximal pole reconstruction using a costo-osteochondral allograft. ¹⁵ The graft is harvested from either the fourth, fifth, or sixth rib at the osteochondral junction. The graft is fixed to the scaphoid with longitudinal K-wires. A review of 47 cases revealed that 85% of the patients rated their outcome as good or excellent and that the majority of patients were able to return to their preinjury vocation without activity modifications. ¹⁵ Additionally, Sandow reported that this technique seems to effectively re-establish the link between the proximal and distal carpal rows without necessitating reconstruction of the SL ligament. He demonstrated no loss of scaphoid alignment or dorsal intercalated segment instability (DISI) deformity developed in his case series.

Bürger et al reported their experience with proximal pole reconstruction using a vascularized MFT based on the transverse branch of the descending geniculate artery. ¹² In their series of 16 cases, 15 healed and 12 of 16 patients had complete pain relief. In addition, average preoperative range of motion and the SL relationship were preserved.

While the feasibility and success of reconstruction of the proximal pole of the scaphoid with a costo-osteochondral or a vascularized MFT osteochondral grafts has been reported, these procedures can be technically challenging and may be associated with donor site morbidity. ^{12 15 18} In addition, they do not directly address reconstruction of the SL ligament, a critical stabilizer of the proximal carpal row. Elhassan et al described a technique for the reconstruction of the proximal pole of the scaphoid with the proximal hamate that also repairs the SL ligament. ¹³ The proximal pole of the hamate is harvested with the volar capitohamate ligament, which is repaired to the remnant of the dorsal SL ligament once this graft is rotated and fixed. Anthropometric assessment of the proximal pole of the hamate in 29 cadavers reported that the majority (69%) of hamates have the appropriate anatomy to serve as a graft for the proximal pole of the scaphoid. ¹⁹ A subsequent cadaveric study demonstrated that resection of the proximal hamate does not adversely affect the kinematics of the wrist. ¹¹ Our study compared reconstruction of the fragmented proximal pole scaphoid nonunion with either a proximal hamate, MFT, or costo-osteochondral graft and found no significant differences in carpal kinematics during wrist flexion–extension and radioulnar deviation between the three reconstructed conditions ( p  > 0.05). Overall, there were also no significant differences in lunocapitate flexion–extension and radioulnar deviation during wrist flexion–extension and wrist radioulnar deviation ( p  > 0.05).

This study has several limitations. First, while our study produced wrist motion by mounting the specimen to an unconstrained X-Y platform and loading wrist flexors and extensors to enable unconstrained carpus movement about natural axes of rotation in an unconstrained manner during simulated wrist motion, this technique lacks the full degrees of freedom present in vivo and during daily kinematics. Nevertheless, we feel that this testing apparatus tried to recreate the in vivo circumstance as best as possible and has been reported in previously. ¹⁴ Second, cadaveric tissue has inherent differences in pliability compared to living tissue and may have deteriorated during testing conditions. In addition, we used a sample of convenience of eight cadavers and the results could be subject to a type II error. The results demonstrated a similar trend, however, between all the cadavers leading us to believe that our findings are accurate.

This study aims to highlight that reconstruction of the proximal pole of the scaphoid with costo-osteochondral, MFT, and proximal hamate grafts resulted in similar carpal kinematics. The proximal hamate graft is harvested from the same operative field, permits repair of the SL ligament, and may be considered as a viable treatment option for unsalvageable proximal pole scaphoid nonunions. Long-term clinical studies are needed to assess the efficacy of this procedure.

Funding Statement

Funding This project was supported by the Mayo Clinic Biomechanics Core and by the Peter Formanek Foundation.