Metaphyseal Core Decompression and Anterograde Fixation for Scaphoid Proximal Pole Fracture Nonunion without Avascular Necrosis

Abstract

Objective  To report the consolidation rate and the results of a series of 22 patients with metaphyseal core decompression of the distal radius and an antegrade compression screw.

Methods  We present a prospective series of patients with scaphoid proximal pole nonunion in whom the presence of intraoperative bleeding was confirmed in both fragments. Patients with displacement, degenerative changes, fragmentation of the proximal pole, cavitation of the focus, loss of height, and necrosis, as well as those with carpal instability, were excluded. The patients were evaluated with X-rays and computed tomography to evaluate their consolidation; their mobility and fist strength were recorded and an analog visual scale (VAS) of pain at rest, pain in activity, subjective functional status, and DASH questionnaire were completed.

Results  Of the 23 patients, 21 accomplished union. The average follow-up was 19 months and the average final range of motion was flexion 86%, extension 85%, radial deviation 79%, ulnar deviation 84%, and grip strength 84%. The average VAS for pain at rest was 1 point, the average VAS for activity pain was 2 points, the average VAS for function was 9 points, and the average DASH score was 8.

Conclusions  Using this simple and reliable technique, we obtained 91% union and very good functional results. Metaphyseal core decompression of the distal radius associated with an antegrade scaphoid screw is a reasonable and effective option for the treatment of proximal pole scaphoid nonunions without avascular necrosis in carefully selected patients.

Level of Evidence  This is Level IV study.

Nonunions of the proximal pole of the scaphoid are a challenge to treat given the intrinsic characteristics of this portion of the bone. First, its small size can hinder fixation, thus preventing rigid stabilization. Second, the fact that it is submerged in synovial fluid and the poor vascular supply because of its retrograde irrigation limit the possibility of achieving union at this level. ¹ Third, the secondary carpal instability that can occur in this type of nonunion would also conspire against the possibilities of consolidation and may evolve to posttraumatic osteoarthritis. ²

Initially, scaphoid proximal pole nonunion was treated by simple resection of the proximal pole or resection and placement of various types of spacers or prostheses, all with poor results. Subsequently, the treatment of this pathology focused on reconstructing the fractured scaphoid through a biological stimulus, both osteogenic and vascular, and providing mechanical stability to the fracture. For this purpose the use of bone grafts, vascularized or not, fixed through pins or screws was incorporated. The two most popular approaches for internal fixation of the scaphoid using self-compressive cannulated screws are the volar and dorsal approaches. The latter facilitates the positioning of the screw in the center of the proximal pole, thus increasing the compression between both fragments without compromising its vascular supply, as long as the capillaries that enter through the “distal” tubercle and into the dorsal capsular ridge of the scaphoid waist are respected. ³ Although the advantage of using vascularized bone grafts in patients with scaphoid proximal pole nonunion with avascular necrosis is clearly reported, this indication would be slightly more controversial in those patients who have a vital proximal pole. This type of surgery demands the need for specific training and adequate equipment that is not always available, constituting a great disadvantage.

In 2001, Illarramendi et al ⁴ reported the technique of metaphyseal core decompression of the distal radius for Kienbock's disease treatment with good final clinical results. Although it has not been demonstrated yet how this simple surgical gesture would act, successful results in the treatment of this disease are well known. It is likely that this procedure will produce a regional vascular change mediated by a cytokine that could improve the vascular flow in the wrist, producing a “local vascular storm” by an as yet unknown mechanism. ⁴ In 2005 a protocol was designed for the treatment of SPPN without carpal collapse or avascular necrosis, which was approved by the ethics committee of our institution. This protocol consisted of a combination of treatments: anterograde scaphoid screw fixation and metaphyseal core decompression of the distal radius for vascular stimulation. Our objective was to understand the healing rate as well as objective and subjective outcomes for a cohort of proximal pole fracture nonunions treated with this novel approach.

Methods

We considered nonunion to any fracture of the proximal pole of the scaphoid with 12 or more weeks of evolution that did not presented signs of union or that had showed evident signs of nonunion in radiographs or axial computed tomography (CT) regardless of the time elapsed since the fracture. We included cases with gap or bone loss present on the preoperative CT scans of 1 mm or less. The Herbert ⁵ and Slade ⁶ classifications were used to define the nonunion type and the Schernberg ⁷ classification ( Fig. 1 ) to define the nonunion location, considering “proximal pole” to be type I and II.

Fig. 1.

Schernberg's classification for scaphoid nonunion.

All patients with scaphoid proximal pole nonunion who did not have radiological or tomographic signs of osteonecrosis and in whom intraoperative bleeding was checked in both fragments with the hemostatic cuff without pressure were included. Only 9 patients were studied with preoperative magnetic resonance imaging (MRI), all showing edema or bone ischemia signal changes but no compatible signs of necrosis.

Patients with degenerative changes, fragmentation or comminution of the proximal pole, cavitation of the focus, and loss of height, and those with carpal instability secondary to nonunion (DISI/VISI) were excluded.

Surgical Technique

All patients were placed in supine position and under local anesthesia the scaphoid was dorsally approached through a 2-cm long incision between the second and fourth extensor compartment and 1 cm distal to the Lister tubercle. A capsulotomy was performed on the ulnar side of the second extensor compartment respecting the dorsal scapholunate ligament.

Those scaphoids presenting with minimal sclerosis and bone resorption at the nonunion interface (3A according to Slade classification) were debrided until punctate bone bleeding was noted with the hemostatic cuff deflated by resecting just the fibrous lamina in the nonunion site. In cases of fibrous nonunion (2A according to Slade classification) the intact cartilaginous envelope was respected.

The wrist was then placed in maximum flexion and ulnar deviation and a cannulated screw was placed in the center of the proximal pole under fluoroscopic control, burying it into subchondral bone and confirming the absence of protrusion at the level of the radioscaphoid joint. In the first cases, 2.3 cannulated cancellous screws (Synthes, Oberdorf, Switzerland) were used, while 2.0 cannulated variable-pitched headless compression screw (Osteomed, Addison, TX) were used in later patients. After saline solution irrigation and hemostasis, the capsule was closed.

In no case we transpose the extensor pollicis longus. In the first cases, the metaphyseal core decompression was performed between the second and fourth dorsal compartment with the same incision that was used to treat the scaphoid, extending 2 cm in the proximal direction. In the latter cases, the approach used was of the same size but located on the lateral aspect of the distal radius, 1 cm proximal to the tip of the styloid of the radius, respecting and avoiding the radial nerve sensory branches. We switched to a lateral radial incision for the metaphyseal core decompression because the lateral cortex of the distal radius is stronger than the dorsal cortex and in this way we would weaken the area less even though this weakening is temporary. In any case, it can be done through a dorsal or lateral exposure. A bone window ∼2 cm long by 0.5 cm wide, beginning 2 cm proximal to the radial styloid, was created with either an osteotome or a bone saw. Through this window, the cancellous metaphyseal bone was impacted without removing it toward the other three cortices, dorsal, ventral, and ulnar, with special caution taken to not compromise them. ^{4 8} After surgery, the wrist was immobilized in a below-elbow cast.

All patients were evaluated at 6 and 8 weeks using without cast X-rays. In cases of possible radiographic consolidation, a CT scan was obtained to confirm the diagnosis. In those cases where radiographic union was not detected wrists were immobilized in a new below-elbow cast for another 6 weeks for union reassessment. In this instance, both radiographs and CT scans were obtained to confirm union.

Bony union was defined as greater than 50% of bone bridging the nonunion site on CT scans. For the subjective evaluation, we used the Spanish validated version of the DASH questionnaire and the visual analog scale from 0 to 10 (where 0 implies complete absence of pain and 10 the maximum) to score their pain at rest, in movement, and function (where 0 implies complete uselessness of the hand and 10 denotes optimal function). Complications and need for reoperation were recorded. Each complication was classified according to the Classification of Surgical Complications ⁹ and defined as any deviation from the expected postoperative evolution that causes pain, functional limitation or required additional treatment (surgical or nonsurgical).

Results

Between 2005 and March 2017, 27 patients with a diagnosis of scaphoid proximal pole nonunion who underwent surgery were identified at our institution. Three of them were excluded due to loss during follow-up, and another patient was excluded due to preoperative radiocarpal osteoarthritis (SNAC 1 stage). A fifth patient had to be removed from the study because of the fragmentation of the proximal pole while being fixed with a cannulated screw. This is considered a complication of the method analyzed and so it is included in the nonunion rate but was excluded from the analysis since its results are the product of a salvage treatment and not of the technique presented here. Of the 22 patients finally included, 4 corresponded to type D1 and 17 type D2 according to the Herbert classification ⁵ according to Slade classification, ⁶ 4 cases corresponded to type 2A and 17 type 3A; and according to the Schernberg classification, ⁹ all patients corresponded to type II ( Table 1 ).

Table 1. Demographic data.

Case no.	Lat	Age (years)	Sex	Herbert's nonunion classification	Slade's nonunion classification	Schernberg's nonunion classification	Nonunion evolution (months)
1	R	30	M	D2	IIIA	II	18
2	L	27	M	D2	IIIA	II	12
3	L	24	M	D2	IIIA	II	4
4	R	33	M	D2	IIIA	II	132
5	L	34	M	D2	IIIA	II	13
6	L	21	F	D2	IIIA	II	6
7	R	30	M	D2	IIIA	II	6
8	R	29	M	D1	IIA	II	N/A
9	L	26	M	D1	IIA	II	36
10	R	20	M	D2	IIIA	II	N/A
11	R	35	M	D2	IIIA	II	36
12	R	27	M	D2	IIIA	II	N/A
13	L	21	M	D1	IIA	II	3
14	L	31	M	D2	IIIA	II	N/A
15	L	36	M	D2	IIIA	II	N/A
16	R	22	M	D2	IIIA	II	12
17	L	30	M	D2	IIIA	II	4
18	L	23	M	D1	IIA	II	8
19	L	22	M	D2	IIIA	II	10
20	L	22	M	D2	IIIA	II	3
21	L	26	M	D2	IIIA	II	8
22	L	29	M	D2	IIIA	II	521
Average		27					49

Abbreviations: F, female; L, Left; Lat, laterality; M, male; R, right.

Twenty-one patients achieved union, confirmed both by radiography and CT scan ( Figs. 2 and 3 ). The average follow-up time was 19 months, and the average range of motion compared with the contralateral side was as follows: flexion 86%, extension 85%, radial deviation 79%, ulnar deviation 84%, and grip strength of 84% ( Table 2 ). According to the visual analog scale, the average rest pain was 1 point, the average activity pain was 2 points, the average function was 9 points, and the average DASH questionnaire result was 8. One patient (case no. 15) evolved unfavorably without showing signs of union but without the presence of pain. Due to personal reasons, this patient postponed his rescue surgery for the next few months. There were only two complications: One patient (case no. 9) presented pain at the screw entry associated with neuropathic pain in the scar located on the lateral aspect of the distal radius through which the metaphyseal core decompression was performed, with complete spontaneous resolution 3 months later. This complication was graded as a type II. The proximal pole outbreak while being fixed with a cannulated screw was considered our second complication and it was graded as a type IIIA.

Fig. 2.

( A, B ) Radiographs and CT scans showing a scaphoid proximal pole nonunion of 4 months of evolution. ( C, D ) One year postoperative radiographs and 4 months postoperative CT scan showing union achievement.

Fig. 3.

A male patient with a history of wrist trauma 10 months before. ( A, B ) Radiographs and CT scans showing a scaphoid proximal pole nonunion. ( C, D ) Six months postoperative radiographs and CT scan showing union achievement.

Table 2. Results.

Case no.	Follow-up (months)	Rest pain VAS	Activity VAS	Function VAS	DASH	Flexion (%)	Extension (%)	Radial deviation (%)	Ulnar deviation (%)	Grip strength (%)
1	15	0	2	9	4	84	82	83	80	83
2	38	0	0	10	2	80	79	90	86	85
3	6	2	5	8	3	75	85	66	75	78
4	54	0	0	10	0	90	78	75	79	80
5	20	2	5	8	17	85	80	85	80	90
6	19	1	1	9	1	90	85	70	82	87
7	12	0	8	7	19	100	89	76	80	133
8	32	0	0	9	0	93	79	78	100	114
9	36	0	3	9,5	5	69	88	77	89	64
10	12	0	3	9	9	114	100	100	90	100
11	32	0	1	6	18	67	44	44	57	76
12	8	0	1	10	0	97	103	100	100	95
13	10	5	3	9	30	97	75	87	92	54
14	5	0	0	9	2	98	100	57	98	63
15	12	0	2	9,5	4,5	88	77	90	93	86
16	14	0	2	8	11	94	100	80	86	78
17	4	0	0	8	2	67	125	100	80	92
18	32	1	5	9	25	75	76	40	67	66
19	14	0	1	8	7	86	85	92	86	80
20	10	1	2	9	5	81	80	83	81	85
21	23	1	0	7	6	80	78	75	73	75
22	15	0	0	7	7	84	90	85	90	80
Average	19	1	2	9	8	86	85	79	84	84

Abbreviation: VAS, visual analog scale.

The range of motions results are described as a percentage of the contralateral side.

Discussion

Several studies have addressed this problem in a specific way and most have focused on necrotic scaphoid proximal pole nonunion. The diversity of methods to define what is considered the scaphoid proximal pole, the multiple treatments proposed and the different diagnostic criteria for union make it difficult to draw conclusions by integrating the available evidence. Merrel et al ¹⁰ performed a meta-analysis conducted on scaphoid nonunion and concluded that the overall success rate of scaphoid proximal pole nonunion treatment is 67%. However, this high failure rate includes patients treated with different techniques, many of which do not include the “biological” contribution.

Several bone grafts options from different donor sites have been reported, all with a high success rate. ^{11 12 13 14} However, this type of surgery not only increases the morbidity of the procedure, but also entails greater technical skills, greater surgical training, an increased economic cost and longer operating room time, all of which are unnecessary factors in patients with a vital proximal pole, without previous surgical failure and with a sufficient proximal pole size to be fixed by a cannulated screw.

Since the concept of metaphyseal core decompression, ⁴ whose form of action has yet not been demonstrated but of evident clinical effectiveness, we added this surgical gesture into the scaphoid proximal pole nonunion treatment. The metaphyseal core decompression acts by increasing the vascularity of the area and generating a form of regional bone repair response, probably mediated by the same substances or cytokines that act in the cascade of events triggered after a fracture has occurred. This would benefit pathologies such as scaphoid proximal pole nonunion or Kienbock where the vascular aspect is compromised. ⁸ This theoretical framework is reinforced by the study of Sherman et al, ¹⁵ where they demonstrated that the metaphyseal core decompression does not alter the forces transmitted through the radiocarpal joint. In their study of radius osteotomies for Kienbock disease treatment, Kam et al ¹⁶ concluded that the mechanical discharge obtained would not be the physiological basis through which the osteotomy acts; rather, they stated that a fibrovascular stimulus secondary to this gesture would be the basis for the success of the osteotomies in the treatment of this pathology. Matsuki et al ¹⁷ reported a retrospective series of 11 consecutive cases treated with nonvascularized bone graft fixed with a Herbert-type screw with a 100% union rate. Using the same type of treatment, DeeMaagd ¹⁸ and Inoue ¹⁹ reported two different series with a union rate of 8 out of 9 patients (89%) and 13 out of 16 patients (81%), respectively. In two patient series with a vital proximal pole treated with an anterograde cannulated screw without bone graft, Krimmer ²⁰ and Herbert ²¹ reported an union rate of 17 out of 23 patients (74%) and 12 out of 16 patients (75%), respectively. In another series, Slade et al ²² reported 15 patients with scaphoid nonunion, all treated with percutaneous internal fixation arthroscopically assisted, and all achieved union. However, its series is not comparable to ours because only 5 out of 15 patients were proximal poles (3 were fibrous nonunions and 2 had minimal sclerosis).

In our study, with scaphoid proximal pole nonunion treated with an anterograde cannulated screw and metaphyseal core decompression, 21 out of 23 patients (91%) achieved union, an incidence that far exceeds the results detailed above and resembles the 100% union reported by Matsuki, ¹⁷ although with a larger number of patients.

The mechanical effectiveness of the scaphoid headless compression screw is determined by 2 factors—screw position and its length. The biomechanical advantages of scaphoid fixation with the screw in central alignment when compared with an eccentric alignment was reported by Mccallister and Trumble. ²³ As they noted, central placement demonstrated 43% greater stiffness and 39% greater load at failure than eccentrically positioned screws. Centrally positioned screws also brings with it the possibility of placing a longer screw with its biomechanical advantages as reported (and recommended) by Dodds et al. ²⁴

In all cases fixation was performed through a dorsal approach. This technical detail is essential to facilitate the central positioning of the osteosynthesis in the scaphoid proximal pole small fragment, thus generating greater compression at the nonunion focus. ^{25 26} Although this fixation route compromises an articular surface, the cartilaginous damage produced by the screw entry site is small and, in the experience of all surgeons taking part in this study, has never produced symptoms while treating cases of scaphoid nonunions or fractures. In 17 cases minimal debridement of the nonunion site was performed, we do not know what the final result would have been if it had not been done in any case. One of our first cases ( Fig. 2 ) was treated using a compression screw with a head buried into subchondral bone and always confirming the absence of protrusion at the radiocarpal joint. In those times (2005) we did not have access to small headless cannulated screws. Actually, we do not recommend this type of implants at all given the availability of smaller headless compression screws. Although our first cases of metaphyseal core decompression were performed through a dorsal approach, we latter switched to a lateral radial incision because the lateral cortex of the distal radius is stronger than the dorsal cortex. Nonetheless, we believe that it can be done by either a dorsal or a lateral approach.

We find this technique to be reliable and simple. It does not require a microscope, microsurgery training, or expensive instruments. It should be noted that our results are in patients with scaphoid proximal pole nonunion with a vital proximal pole and without carpal collapse. We ignored the evolution of this technique in patients with proximal necrosis or with carpal instability since they were not the subject of this study.

The present study has several limitations. First, we had a relatively low number of patients analyzed. Nevertheless, the very low incidence of such injuries may justify the limited number of patients enrolled. Second, the influence of the entire process of nonunion and its consolidation toward an eventual degenerative process of the wrist has not been studied in the long term. In the near future, with a longer follow-up period, we expect to be able to answer any questions that arise and those which have not been answered in the present work. Finally, despite we had satisfactory clinical and radiological outcomes using the radius metaphyseal core decompression for the treatment of carpus pathology, we acknowledge that there is a lack of basic sciences studies supporting our revascularization hypothesis.

Despite these limitations, we had satisfactory clinical and radiological outcomes using scaphoid screw fixation along with and metaphyseal core decompression of the distal radius for the treatment of selected proximal pole fracture nonunions of the scaphoid.