Scaphoid non-union: Factors affecting the functional outcome of open reduction and wedge grafting with herbert screw fixation

SCAPHOTD NON-UNION: FACTORS AFFECTING THE FUNCTIONAL OUTCOME OF OPEN REDUCTION AND WEDGE GRAFTING WITH HERBERT SCREW FIXATION R. NAKAMURA,

E. HORII, K. WATANABE, K. TSUNODA

and T. MIURA

From the Division of Hand Surgery, Department of Orthopaedic Surgery, Nagoya University School of Medicine, Japan

50 patients with scaphoid non-union were treated by open reduction, anterior wedge bone grafting and internal fixation using the Herbert screw. I&a-operative image intensifier control enabled us to insert the screw into the scaphoid accurately. An excellent or good functional outcome was less likely when more than 5 years had elapsed since injury, the non-union was in the proximal third, when sclerosis of the proximal fragment was present, and when reduction of carpal and scaphoid deformity was unsatisfactory. These four factors are believed to be the primary determinants affecting the functional results of the surgical treatment of scaphoid non-union, even when bony union is achieved.

Journal of Hand Surgery (British and European Volume, 1993) 18B: 219-224 treated by early cast immobilization. Details of the patients are summarized in Table 1. Although the mean age in group C at the time of surgery was significantly higher than in the other two groups, the mean age at the time of injury was the same. The most frequent site of fracture was the waist of the scaphoid in all three groups, and 38 (76%) wrists developed dorsal intercalated segment instability (DISI) deformity (Linscheid et al, 1972), defined as a difference in the radiolunate angle (Sarrafian et al, 1977) of more than 10” between the affected and unaffected wrists. The carpal height ratio (Youm et al, 1978) was more than 0.03 less on the affected than unaffected side in 12 (24%) wrists, and 15 (30%) patients with non-union had scaphoid shortening (Fernandez, 1984) of 2 mm or more. Postoperative follow-up ranged from 1 to 6 years, without significant differences between groups.

5 to 10% of acute scaphoid fractures proceed to nonunion after immobilization in a cast (London, 1961; Eddeland et al, 1975; Leslie and Dickson, 1981). Patients with non-union occasionally seek treatment months or years after injury because the symptoms are minimal, so the percentage with non-union may be higher. Surgery is the current treatment of choice for scaphoid non-union. In the absence of extensive osteoarthrosis, our approach is open reduction, anterior wedge-shaped bone graft (Fisk, 1970; 1984) and internal fixation using a Herbert screw (Herbert and Fisher, 1984), as we believe that anatomical reduction is the technique most likely to yield satisfactory functional results. This paper describes our results of surgical treatment of scaphoid non-union and identifies those factors having the greatest effect on functional outcome. MATERIAL

AND METHODS Operative technique

Between 1985 and 1990, 64 patients with scaphoid nonunion, excluding those associated with perilunate dislocation, who presented more than 3 months after the injury, were seen at Nagoya University Hospital. Five patients with extensive osteoarthritic changes of the wrist were treated by resection of the proximal fragment and tendon ball insertion or radial styloidectomy (Bentzon and Madson, 1946). Three patients refused surgical treatment. 56 patients underwent internal fixation using the Herbert screw with bone grafting. Six patients were lost to follow-up within 1 year of radiographic union. The 50 patients who were followed for more than 1 year were included in the study. 18 (group A) underwent surgery from 3 to 10 months after injury, 20 (group B) from 1 to 4 years after injury and 12 (group C) from 5 to 30 years after injury. 12 patients (six each in groups A and B) had undergone cast immobilization previously. The diagnosis was delayed in eight patients (five in group A and three in group B). 30 patients sought treatment a long time after injury. Thus three out of four patients with scaphoid non-union had not been

The anterior approach (Herbert and Fisher, 1984) was used. The distal and anterior surfaces of the scaphoid were exposed, and the fibrous tissue between the two fragments was excised. Among patients with proximal third non-union, three required a dorsal approach and retrograde insertion of a Herbert screw (Herbert and Fisher, 1984; DeMaagd and Engher, 1989), and one other patient with proximal third non-union and DISI deformity underwent retrograde insertion of a Herbert screw and anterior wedge grafting, using a combined anterior and dorsal approach. A 1.5 mm Kirschner wire was inserted from the dorsum of the wrist into the lunate to facilitate reduction under image intensifier control in 38 cases of scaphoid non-union with DISI deformity. The lunate was brought into the palmarflexed position by the Kirschner wire, and the DISI deformity was reduced (Nakamura et al, 1987; 1991). Cortico-cancellous bone graft was obtained from the iliac crest or the distal radius and was sculptured to match the anterior bony defect in the scaphoid which developed after reduction. In cases of scaphoid non219

220

THE JOURNAL OF HAND SURGERY VOL. 18B No. 2 APRIL 1993

Table l-Patients Patient group

No. of patients

with scaphoid non-union undergoing Herbert Screw insertion

Sex male

female

Age at surgery (years)

A

18

16

2

B

20

17

3

C

12

9

3

Total

50

42

8

13-26 (19,4) 13-39 (20 i 8) 25-57 (36+ 11) 13-57 (24+ 10)

Interval between injury and operation (mos.) 3-11 (7f2) 12-59

distal %

Fracture site waist proximal %

No. of patients with DISI deformity*

No. of patients with decreased carpal height ratio?

No. of patients with shortening scaphoidt”f

1

15

2

14

6

6

4

15

1

16

4

6

60-360 (336+ 108)

2

6

4

8

2

3

3-360 (52+88)

7

36

7

38

12

15

(20 It 9)

Follow-up time in months

12-48 (20 + 9) 12-72 (20,15) 12-48 (21& 13) 12-72 (2Ok 12)

Data are presented as the range and mean k SD, where appropriate. *DISI deformity was defined as a difference of more than 10” in the radiolunate angle or scapholunate angle between the affected and unaffected wrists tPatients with a carpal height ratio of 0.03 or less than the unaffected wrist. TtPatients with a scaphoid length 2 mm or more shorter in the affected than the unaffected wrist.

union without DISI deformity, cortico-cancellous bone graft was inserted when the bony defect was large, and multiple chip grafts were used when the defect between two fragments was small. Internal fixation was accomplished using a Herbert screw under image intensifier control, and the Kirschner wire was removed. While the screw was being inserted, the bone graft was fixed by a small impactor to maintain the position of the graft. Cast immobilization was continued for 4 to 8 weeks after surgery. Pre- and post-operative X-ray examinations included the standard postero-anterior and lateral views, and a motion study. X-rays of the contralateral wrist were also obtained pre-operatively. Results were evaluated according to the rate of union, wrist pain, range of flexion and extension and grip strength. Overall results were graded on the wrist evaluation scale originally proposed by Green and O’Brien (1978) and modified by Cooney et al (1987). A score of 90 or above was rated excellent, 80 to 89 good, 65 to 79 fair and below 65 was poor. Post-operative results were analyzed as a function of the length of time since injury, the site of the nonunion, pre- and post-operative carpal alignment, the presence or absence of sclerosis of the proximal fragment and extensive sclerosis or cyst formation at the fracture site. Herbert screw insertion was satisfactory in all but three patients. One screw was too long, protruded proximally and required removal after bony union. Two screws failed to fix the scaphoid rigidly and required the insertion of additional Kirschner wires. RESULTS Union was obtained in 47 out of 50 patients (94%). Three patients in group C showed persistent non-union postoperatively, and all of their pre-operative X-rays showed sclerotic changes in the proximal fragment. The screw head became loose in two of these three patients a few months after the surgery.

Post-operatively, the radio-lunate angle, scapholunate angle, carpal height ratio, and scaphoid length improved in all patients who underwent DISI reduction (Figs 1 and 2). However, reduction was unsatisfactory in six patients (more than 20” lower radiolunate angle or more than 20” higher scapholunate angle than the contralateral wrist). Sclerosis of the proximal fragment was seen pre-operatively in 11 non-unions, five of which showed mild collapse of the proximal fragment. Sclerotic changes in the proximal fragment improved in five patients and deteriorated into collapse of the proximal fragment in two post-operatively (Fig 2). Mild radiocarpal osteoarthritis was seen in only one patient with sclerosis of the proximal fragment pre-operatively and another three patients developed radiocarpal osteoarthritis post-operatively. Clinical results were summarized in Table 2. Group C patients showed the least improvement of wrist pain, range of wrist flexion and extension, and grip strength post-operatively. The intergroup difference was significant for wrist range of motion between group A and C (t test, 0.02 > P > 0.01) and for increase in group strength between group A and C (t test 0.05 > P> 0.02). Excellent or good results were obtained 14 of 18 patients in group A, 14 of 20 patients in group B and one of 12 patients of group C (Table 2). The difference in the frequency of an excellent or good outcome was significant between group A and B and group C (chisquare test with correction for continuity, P < 0.001 and 0.02 < P < 0.05 respectively). There was significantly fewer excellent or good outcome in cases of proximal third non-union and in non-union with a sclerotic proximal fragment (chi-square test with correction for continuity, 0.02~ PcO.05) (Fig 1). Of six patients with an unsatisfactory reduction, two failed to unite; clinically these move one good and five fair results. The quality of reduction was believed to affect the results although a statistical difference was not confirmed (chi-square test with correction for continuity: 0.05 < P < 0.1). Pre-

SCAPHOID NON-UNION

Fig 1

Scaphoid waist non-union 5 months after injury. Pre-operative X-rays (a and b) showing cystic change at the site of non-union and DISI deformity of the wrist. Post-operative X-rays (c and d) obtained 20 months after the surgery show bony union with satisfactory carpal and scaphoid alignment and an excellent functional result.

operative sclerosis or cystic change at the site of nonunion (Fig 1) and pre-operative DISI deformity did not significantly affect the outcome (Table 3). DISCUSSION

The use of a screw for internal fixation in scaphoid fractures, including non-union, was first advocated by McLaughlin (1954). However, the technical difficulty of inserting a screw into the complicated three-dimensional anatomy of the scaphoid inhibited the adoption of this technique. Recent refinements in the instruments, screw and technique have led to improved rates of union with a short period of immobilization, and screw fixation has come into common use (Herbert and Fisher, 1984; Leyshon et al, 1984; Brostrijm et al, 1986; Bunker et al, 1987; Ford et al, 1987; Nakamura et al, 1987; Pring

et al, 1987; Adams et al, 1988; Cooney et al, 1988; Manske et al, 1988; Warren-Smith and Barton, 1988; Fernandez, 1990; Radford et al, 1990; Wozasek and Moser, 1991). However, the success rate is still less than 90% in most of these reports, which is no better than the rate of union achieved by the Matti-Russe technique (Russe, 1960; Muldler, 1968; Green, 198.5; Stark et al, 1988; Hooning Van Dupvenbode et al, 1991), although Warren-Smith and Barton (1988) found that Herbert screw fixation gave a higher success rate in achieving bony union with better functional results. The technical difficulty of screw insertion has remained a major criticism of screw fixation. As reported by Fernandez (1990), the most common cause of failure is poor fixation technique or failure to perform bone grafting. Our results have confirmed this, and we believe that our success in obtaining bony union was based on reduction

222

Fig 2

THE JOURNAL OF HAND SURGERY

Scaphoid waist non-union 10 years after injury. Pre-operative X-rays (a and b) showing sclerosis of the proximal fragment with cystic changes and a DISI deformity of the wrist. Post-operative X-rays (c and d) obtained 20 months after insertion of a Herbert screw show collapse of the proximal fragment. In spite of bony union and satisfactory reduction of the DISI deformity, the functional result was only fair.

Table 2-Surgical

results of Herbert Screw fixation for scaphoid non-union Persistent non-union

_

Wrist pain I!z

+

A B C

0 0 3

11 12 1

6 I 11

1 1 0

70&18/90*13 81k 9/88&11 77 & 14178f 15

71*25/91*13 81+25/95*11 79* 15/84& 12

Total

3

24

24

2

76+ 15/86& 13

78f20/91i

Group

Data are

VOL. 18B No. 2 APRIL 1993

presented as the

mean f SD, where

appropriate.

Range of wrist flexion-extension (percentage of) unaffected side) preop./postop.

Grip strength percentage of unaffected side preop./postop.

12

Excellent

Good

Fair

Poor

6 8 0

8 6 1

3 6 11

1 0 0

14

15

20

1

SCAPHOID

223

NON-UNION

Table 3-Factors

affecting the outcome of Herbert Screw fixation for scapboid non-union

Outcome distal third

Fracture site waist proximal third

Pre-operative proximal fragment sclerosis (+I i-)

Extensive pre-operative sclerosis or cyst formation at non-union site (+)

C-i

Pre-operative DISI* deformity (+I

C-1

Reduction satisfactory unsatisfactory

Excellent Good Fair Poor

1 3 3 0

13 11 I1 1

0 1 6 0

2 1 8 0

12 14 12 1

5 4 4 1

9 11 16 0

11 11 15 1

3 4 5 0

11 10 10 1

0 1 5 0

Total

I

36

7

11

39

14

36

38

12

32

6

*DISI: dorsal intercalated segment instability

of DISI deformity, anterior wedge grafting and accurate screw positioning, as verified by image intensifier. Reduction of DISI simultaneously corrects humpback deformity, yields a straighter scaphoid, and facilitates screw insertion. Wedge grafting preserves correction of the humpback deformity and provides an adequate contact area for bony union. Even after anatomical alignment, the proper application of the jig is difficult and image intensifier control is advisable (Pring et al, 1987). Our approach to the surgical treatment of scaphoid non-union is based on the principle that restoration of normal anatomy leads to better functional results. We have found that bony union is a prerequisite for improved wrist function, but does not assure an excellent or good outcome. The interval between injury and operation, the location of the fracture site and the presence of sclerotic change in the proximal fragment have a significant effect on the surgical result. The influence of these factors was most evident in group C. Radford et al (1990) pointed out that delay in surgery seems to be the major factor predisposing to persistent delayed union or non-union. We agree that satisfactory functional results can be expected in patients treated less than 5 years after injury. For non-union with a sclerotic proximal fragment, screw fixation has been recommended by Fernandez (1990) but has been questioned by Cooney et al (1988). Our results show that bony union can be obtained in most patients (8 of 11) even when the proximal fragment is sclerotic, but functional improvement is less than when sclerosis is not present. To improve function, post-operative traction by an external fixator or some other measure to protect or decompress the proximal fragment may be required. Although Cooney et al (1988) have recommended the Matti-Russe technique for non-union with a sclerotic proximal fragment, further studies comparing the two techniques are required before a final conclusion can be made. Unsatisfactory reduction yields unsatisfactory results. This was also the conclusion of Amadio et al (1989), who found that scaphoid malunion was occasionally symptomatic.

In conclusion, functional improvement is to be expected if non-union is treated within 5 years of injury, but is unlikely when more than 5 years have elapsed, when there is sclerosis of the proximal fragment, when the site of non-union is in the proximal third of the scaphoid, or when reduction of the carpal and scaphoid deformity is unsatisfactory. References ADAMS, B. D., BLAIR, W. F., REAGAN, D. S. and GRUNDBERG, A. B. (198X). Technical factors related to Herbert screw fixation. Journal of Hand Surgery, 13A: 6: 893-899. AMADIO, P. C., BERQUIST, T. H., SMITH, D. K., ILSTRUP, D. M., COONEY, W. P. and LINSCHEID, R. L. (1989). Scaphoid malunion. Journal of Hand Surgery, 14A: 4: 679-687. BENTZON, P. G. K. and RANDL0V-MADSEN, A. (1946). On fracture of the carpal scaphoid: A method of operative treatment of inveterate fractures. Acta Orthopaedica Scandinavica, 16: 30-39. BROSTROM, L.-A., STARK, A. and SVARTENGREN, G. (1986). Non-union of the scaphoid treated with styloidectomy and compression screw fixation. Scandinavian Journal of Plastic and Reconstructive Surgery, 20: 289-291. BUNKER, T. D., McNAMEE, P. B. and SCOTT, T. D. (1987). The Herbert screw for scaphoid fractures: A multicentre study. Journal of Bone and Joint Surgery, 69B: 4: 631-634. COONEY, W. P., BUSSEY, R., DOBYNS, J. H. and LINSCHEID, R. L. (1987). Difficult wrist fractures: Perilunate fracture-dislocations of the wrist. Clinical Orthopaedics and Related Research, 214: 136-147. COONEY, W. P., LINSCHEID, R. L., DOBYNS, J. H. and WOOD, M. B. (1988). Scaphoid nonunion: Role of anterior interpositional bone grafts. Journal of Hand Surgery, 13A: 5: 635-650. DEMAAGD, R. L. and ENGBER, W. D. (1989). Retrograde Herbert screw fixation for the treatment of proximal pole scaphoid nonunions. Journal of Hand Surgery, 14A: 6: 996-1003. EDDELAND, A., EIKEN, O., HELLGREN, E. and OHLSSON, N. M. (1975). Fracture of the scaphoid. Scandinavian Journal of Plastic and Reconstructive Surgery, 9: 234-239. FERNANDEZ, D. L. (1984). A technique for anterior wedge-shaped grafts for scaphoid nonunions with carpal instability. Journal of Hand Surgery, 9A: 5: 733-737. FERNANDEZ, D. L. (1990). Anterior bone grafting and conventional lag screw fixation to treat scaphoid nonunions. Journal of Hand Surgery, 15A: 1: 140-147. FISK, G. R. (1970). Carpal instability and the fractured scaphoid. Annals of the Royal College of Surgeons of England, 46: 63-76. FISK, G. R. (1984). The Wrist. Journal of Bone and Joint Surgery, 66B: 3: 396-407. FORD, D. J., KHOURY, G., EL-HADIDI, S., LUNN, P. G. and BURKE, F. D. (1987). The Herbert screw for fractures of the scaphoid. Journal of Bone and Joint Surgery, 69B: 1: 124-127. GREEN, D. P. and O’BRIEN, E. T. (1978). Open reduction of carpal dislocations: Indications and operative techniques. Journal of Hand Surgery, 3: 3: 250-265. GREEN, D. P. (1985). Th e e ffec t o f avascular necrosis on Ruse bone grafting for scaphoid nonunion. Journal of Hand Surgery, 10A: 5: 597-605. HERBERT, T. J. and FISHER, W. E. (1984). Management of the fractured

224 scaphoid using a new bone screw. Journal of Bone and Joint Surgery, 66B: 1: 114-123. HOONING VAN DUYVENBODE, J. F. F., KEIJSER, L. C. M., HAUET, E. J., OBERMANN, W. R. and ROZING, P. M. (1991). Pseudoarthrosis of the Scaphoid Treated by the Matti-Russe operation: A long-term review of 77 cases. Journal of Bone and Joint Surgery, 73B: 4: 603-606. LESLIE, I. J. and DICKSON, R. A. (1981). The fractured carpal scaphoid: Natural history and factors influencing outcome. Journal of Bone and Joint Surgery, 63B: 2: 225-230. LEYSHON, A., IRELAND, J. and TRICKEY, E. L. (1984). The treatment of delayed union and non-union of the carpal scaphoid by screw fixation. Journal of Bone and Joint Surgery, 66B: 1: 124-127. LINSCHEID, R. L., DOBYNS, J. H., BEABOUT, J. W. and BRYAN, R. S. (1972). Traumatic instability of the wrist: Diagnosis, classification and pathomechanics. Journal of Bone and Joint Surgery, 54A: 8: 1612-1632. LONDON, P. S. (1961). The broken scaphoid bone: The case against pessimism. Journal of Bone and Joint Surgery, 43B: 2: 237-244. MANSKE, P. R., MCCARTHY, J. A. and STRECKER, W. B. (1988). Use of the Herbert bone screw for scaphoid nonunions. Orthopedics, 11: 12: 1653-1661. MCLAUGHLIN, H. L. (1954). Fracture of the carpal navicular (scaphoid) bone: some observations based on treatment of open induction and internal fixation. Journal of Bone and Joint Surgery, 36A: 4: 765-774. MULDER, J. D. (1968). The results of 100 cases of pseudarthrosis in the scaphoid bone treated by the Matti-Ruse operation. Journal of Bone and Joint Surgery, 50B: 1: 110-115. NAKAMURA, R., HORI, M., HORII, E. and MIURA, T. (1987). Reduction of the scaphoid fracture with DISI alignment. Journal of Hand Surgery, 12A: 6: 1000-1005. NAKAMURA, R., IMAEDA, T., TSUGE, S. and WATANABE, K. (1991). Scaphoid non-union with DISI deformity: A survey of clinical cases with

THE JOURNAL

OF HAND

SURGERY

VOL. 18B No. 2 APRIL

1993

special reference to ligamentous injury. Journal of Hand Surgery, 16B: 2: 156-161. PRING, D. J., HARTLEY, E. B. and WILLIAMS, D. J. (1987). Scaphoid osteosynthesis: Early experience with the Herbert bone screw. Journal of Hand Surgery, 12B: 1: 46-49. RADFORD, P. J., MATTHEWSON, M. H. and MEGGITT, B. F. (1990). The Herbert screw for delayed and non-union of scaphoid fractures: A review of fifty cases. Journal of Hand Surgery, 15B: 4: 455-459. RUSSE, 0. (1960). Fracture of the carpal navicular: Diagnosis, non-operative treatment, and operative treatment. Journal of Bone and Joint Surgery, 42A: 5: 759-768. SARRAFIAN, S. K., MELAMED, J. L. and GOSHGARIAN, G. M. (1977). Study of wrist motion in flexion and extension. Clinical Orthopaedics and Related Research, 126: 153-159. STARK, H. H., RICKARD, T. A., ZEMEL, N. P. and ASHWORTH, C. R. (1988). Treatment of ununited fractures of the scaphoid by iliac bone grafts and Kirschner-wire fixation. Journal of Bone and Joint Suraerv. - ,; 70A: 7: 982-991, WARREN-SMITH, C. D. and BARTON, N. J. (1988). Non-union of the scaphoid: Ruse graft vs Herbert screw. Journal of Hand Surgery, 13B: 1: 83-86. WOZASEK, G.-E. and MOSER, K.-D. (1991). Percutaneous screw fixation for fractures of the scaphoid. Journal of Bone and Joint Surgery, 73B: 1: 138-142. YOUM, Y., MCMURTY, R. Y., FLATT, A. E. and GILLESPIE, T. E. (1978). Kinematics of the wrist I. An experimental study of radialulnar deviation and flexion-extension. Journal of Bone and Joint Surgery, 60A: 4: 423-431,

Accepted: 9 June 1992 Ryogo Nakamura, MD, Branch Hospital of Nagoya Daikominami, Higashiku, Nagoya 461, Japan. 0

1993 The British

Society

for Surgery

of the Hand

University

School

of Medicine,

I-1-20