grams¯ V. Acl~.. Tt:et it ,.:, asndiagllosticell

State Un

nai Hospital

gen aspects fuetral joint.

orm trian.ular951riquetral.

ronunion of the scaphoid:bone grafting

Analysis of the results

nonunion was treated by bone-grafting techniques in 86 patients with 90 scaphoid fractures. Theinlay" (Russe) technique had union in 38 of 44 fractures (86%); the dorsal inlay (Mat#) technique

r22 fractures (91%); the dorsal peg graft (Murray) in nine of 18fractures (50%); and compressionosteosynthesis in one of six cases (17%). Fourteen patients had secondary arthritis associated with

union. Union was more frequent (85%) in 54 undisplaced nonunions than in 36 displaced(65%). Complications included avascular necrosis, secondary arthrosis, and persistentAvascular necrosis in 13 nonunions was treated successfully by inlay grafts in 11, and

~erative avascular necrosis (four nonunions) was resolved after fracture healing in two. The fourwith failed results had the dorsal peg grafting technique. Q( 22 persistent nonunions, 16 had

m...~-v arthritis and four had avascular necrosis. The rate of union was increased when Kirschnerof unstable nonunions was performed. The dorsal and volar inlay techniques had consistently

rares of union with fewer complications than did the dorsal peg grafting technique.

lliam P. Cooney III, M.D., James H. Dobyns, M.D., andL. Linscheid, M.D., Rochester, Minn.

ne. Hand Chit

eport. J

ce, 1959,

of the scaphoid is dependent on a number ofuisites: early diagnosis, bone apposition, ade-blood supply, and fracture immobilization.~-a

location of the fracture, the degree of displacementof the fracture fragments, and the length of

.~-. ~,(time) are4important factors that contribute

~,:-.:i~racture healing,a, When fracture union is not ob-.:)!’~ned, bone grafting has been recommended to help

these prerequisites. However, the best method ofical treatment is controversial with respect to the

ue of bone grafting, the need for internal fixa-and the effect of the surgical approach on the

~hoid’s blood supply. Many surgical proceduresbeen advocated for the treatment of the scaphoid

which suggests that difficulties exist in

~ this problem and that no method has been uni-:<~:-!~ersally:’~ accepted or successful.’~ We analyzed the re-" ii~!/~llts,.___~ of treatment of scaphoid nonunion by bone-!:?~~ting techniques to judge whether one methodof;~eatment~ is preferable to the others and to determine in.:~hat manner the above-related factors contribute to;i ~ ~aphoid union¯

the Section of Hand Surgery, Department of Orthopedics,Mayo Clinic and Mayo Foundation, Rochester, Minn.

for publication July 3, 1979.

requests: William P. Cooney III, M.D., Section of HandSurgery, Department of Orthopedics, Mayo Clinic and MayoFoundation, Rochester, MN 55901.

Methods and material

The material reviewed in this study comes from atotal series of 946 wrist injuries with 435 scaphoid frac-tures that were seen at our institution between 1961 and1976. In that group of patients, there were 110scaphoid nonunions in 106 patients. From this group of106 patients, we excluded 20 patients who were treatedby techniques other than bone grafting. This includedarthrodesis (six patients), silicone rubber replacement(four patients), radial styloidectomy (seven patie.n_.ts),and soft-tissue arthroplasty (Bentzon~) (three patients).Patients who had fractures of the scaphoid associatedwith delayed union or nonunion of less than 6 monthsin duration often were treated by cast immobilizationalone and were not included in the series. Patients withasymptomatic nonunions in whom the diagnosis wasmade on roentgenographic evidence alone were simi-larly excluded.

The series involved 86 patients with 90 scaphoidnonunions. Clinical examination and roentgenogramswere obtained on 80 of the 86 patients over a median of5 years (range, 12 to 163 months) from the time of theirinitial surgical treatment. One patient died, and fivewere lost to follow-up. These six patients had sufficientclinical information recorded and were seen at least 1year after the operative procedure and so were includedin part. of this studyl However, they were not includedin the long-term results. Twenty of the patients, with

1/80/040343+ 12501.20/0 © 1980 American Society for Surgery of the Hand THE JOURNAL OF HAND SURGERY343

344 Cooney et al.The Journal

HAND SURGEII

Fig. 1. Stable nonunion should be differentiated from unstable nonunions. Anteroposterior (A),lateral (B), and oblique (C) views may show comminution and displacement of greater than (traction views).

clinical follow-up of less than 18 months, responded toquestionnaires and indicated the amount of wrist pain,grip weakness, and range of motion they experienced.

The mean age of the patients in this series was 23years, with a range of 15 to 72 years. Ninety percentwere men; 54% had the dominant hand involved. Atdiagnosis of the nonunion, from 4 months to 30 yearshad elapsed from the date of the original injury. Mostof the patients, however, presented with symptoms be-tween 6 months and 3 years from the time of injury,and 90% had presented by 5 years. Twenty-seven pa-tients reported no initial treatment of their wrist injury,six had splints applied, 39 had cast support (25 short-arm, nine long-arm, and five thumb-spica casts), and14 did not remember the initial injury or treatment. Theca.use of the fracture was a fall in 16 patients, a motorvehicle accident in 17 patients, sports-related injury in32 patients, and industrial-related injury in five pa-tients; the mechanism was unknown in I6 patients.

-The presenting complaint and indication for opera-tion was pain alone in 43 patients (50%), pain associ-ated with loss of motion in 20 patients (23%), and pain,loss of motion, and a weak grip in 14 patients (16%).Six patients (7%) were totally disabled from any work.Three patients (4%) had a median neuropathy at thewrist associated with pain, restricted motion, and re-duced grip. Grip strength was 27 kg (median) in theinjured wrist and 44.5 kg in the uninjured wrist, andwrist motion was reduced in the injured wrist with amedian range of 42° of dorsiflexion and 52° of volarflexion at the time of initial assessment.

Radiographic study identified the location of the

nonunion and the direction of the fracture line andclassified the nonunions into two types: stable undis-placed and unstable displaced. The scaphoid view withtraction (Fig. 1, A, B, and C), wrist motion series (Fig.I, D, E, F, and G), and trispiral tomography (Fig. H) assessed the degree of fracture displacement andinstability. The fracture was judged to be displacedthere was greater than 45° of scapholunate angulation(DISI deformity) on the lateral view (Fig. 1, I) Of

one-third,i~l~!90 fractures, 32 (36%) were in the proximal56 (62%) in the mid one-third, and two (2%) in distal one-third (Fig. 2). There were 38 displaced

and four fractureii~i~~itures, 48 stable undisplaced fractures,dislocations--the latter being determined from roent2genograms of the initial injury. The fracturg-linehorizontal in 43 patients, transverse in 32, spiral inthree, and comminuted in eight. The diagnosisnonunion was made from the plane film in 90%,motion films (radial and ulnar deviation views) in Iand from trispiral tomograms in 4%.

Treatment of scaphoid nonunions by bonealone or combined with internal fixation (Fig. 3)performed using four techniques by eightSubdivided into the four categories, they were (1)bone graft (Russe*a technique), 44 fractures; (2)peg graft (Murrayt° technique), 18 fractures; (3)inlay graft (Matti, 9 Barnard-Stubbins6 technique,modified by Lipscomb and Dobyns*4), 22 fractures(4) compression osteosynthesis (McLaughlin techiniqueTM ~6), six fractures. No patient was treateddrilling of the fracture nonunion alone. Of the Russe,iprocedures, internal fixation was performed in 20 of the

[ournal 5. No. 41980 Nonunion of scaphoid 345

Fig. 1, cont’d. Radioulnar deviation views can indicate a stable nonunion without motion at thenonunion site (D and E), whereas an unstable nonunion will have motion (F and G) at the nonunionsite. A tomogram (H) of the nonunion site may show displacement, and the lateral view (I) demonstrate dorsal rotation of lunate [lunocapitate angulation of greater than I0° and scapholunate(S-L) angle of greater than 45°] in the unstable nonunion.

line and~le undis-view withcries ~ Fig.y (Fig. 1,~’ment andisplaced ifangulationI). Of the.one-thir , ~~:,_!%) in the i~:i;

ur fracture ~;~:.

e line was IiI, spiral in ~i;..:agnosis of’~2:.

. :(~~ fracture90%, fron~ nonunions. In all of the dorsal inlay grafts,ws) in 6% or more Kirschner wires were utilized, and in four

the 18 peg grafts, one or more Kirschner wires wereae grafting ~:ig. 3)wasl,

surgeonsre

(2) dors~(3)

hnique,~ctures; and;hlintreatedthe Russe

in 20 of the

indication for surgical intervention was primar-related to subjective complaints of pain, restricted

and loss of strength; to objective findings oftenderness, pain at the extremes of motion,

decreased grip; to roentgenographic nonunion withcystic changes, and motion at the fracture site

4). Of the 86 patients, 53 (62%) had nonunion12 (14%) had nonunion and associated avascular

18 (21%) had nonunion and radioscaphoidand four (4%) had posttraumatic instability

from fracture dislocations (one within 4 months of the

Number ofLocation % UnionFractures

Distal I/3 2 I00Middle I/3 56 80

~ Proximal 1/3 32 64

Number ofDisplacement % UnionFractures

,/’~--- Stable 48 85

l~42 65

~Onstable,.,.~ ~

Fig. 2. Union of the scaphoid after bone grafting is influencedsignificantly by both the location of the fracture and thearnount of displacement.

346 Cooney et al.The Journal

HAND SURGE

Fig. 3. Scaphoid nonunions were treated by four operative techniques utilizing cortical or cor-ticocancellous bone grafts with internal fixation for unstable nonunions: A, volar inlay (Russe); dorsal peg (Murray); C, dorsal inlay and radiosty]~oidectomy; D, compression screw and matchstickgraft.

Table I. Results of bone grafting of initial scaphoidnonunions

UnionNo. of Non- Un-

Grafting technique cases No. % union certain

Volar inlay(Russe) 44 38 86 4 2

Dorsal inlay 5::(Matti, Barnard ....Stubbins modi-fication)* 22 20 91 2 0

Dorsal peg(Murray) 18 9 50 6 3 .....

ASIF screwosteosynthesis 6 1 l~7 ~ 0

Total 90 68 76 17 --- 5

*Dorsal approach modified by Lipscomb and Dobyns~4 by using radiosty-loidectomy and inlay-grafting technique of Matti.

Fig. 4, Cystic resorption, sclerosis, and round edges indicatean early established nonunion. When these roentgenographicfindings are associated with appropriate clinical signs orsymptoms, bone grafting of nonunion is indicated.

injury and three at more than 12 months from the timeof the initial injury).

Immobilization after the bone-grafting procedure

was applied in all patients; a long-arm cast for a medianof 7 weeks (range, 4 to 20 weeks) and a short-arm castfor a median of 6 weeks (3 to 24 weeks). A splint wasprescribed for protected mobilization in 48 of the 86patients. Total immobilization time was 17 weeks,ranging from 11 weeks to 50 weeks.

Table II. Complications associated with scaphoidnonunion

TechniqueNo. of I Avascular Arthrosiscases necrosis I with union

Volar inlay(Russe) 44 1 * 4 6

Dorsal inlay(Barnard-Stubbinsmodification) 22 1 * 4 2

Peg graft(Murray) 18 2~t 5 9

ASIF screwosteosynthesis .~6

0__0_-_~1 ~

Total 90 4 14 22

*Uniort occurred after inlay graft and the avascuiar necrosis resolved.t Nonunion and avascular necrosis after two dorsal peg grafts.

NO.

1980Nonunion of scaphoid 347

":~nunion after bone grafting. A, Delay of 6 monthsdiagi~.:~:ais of scaphoid nonunion. There is an absence of

or cystic changes at fracture site. B, After cast im-ilization for 3 months and 9 months from the original

(case 1), union appeared to be present. C, However, 1 months from the time of the original injury,ident. D, A Russe bone graft was performed.

union rate of the 90 scaphoid nonunions was(’Table I). The volar inlay (Russe) technique

in union in 38 (86%) of 44 cases, with fourand two uncertain. The dorsal inlay technique

loidectomy had union in 20 of 22 cases~. The dorsal peg graft (Murray technique) had

nine of 18 cases (50%), with six definiteand three uncertain unions. Five of the six

treated by compression screw osteosynthesis.) were ununited. All initially provided ex-

scaphoid stability; but by 2, 2.5, and 4 years ofscrew loosening and backing out were noted

of the five patients (four of the six fractures).two of these patients had instability with

were treated by dorsal inlay bone graft. Thetures (both in one patient) were not clini-

union was determined by both clinical and9hic examination. Fracture unionin the

results could not be determined by roentgen-that included trispiral tomography, and these

Fig. 5. E, At 14 months complete resorption of the bone grafthad occurred, with symptomatic nonunion. F, Unstablenonunioff’with dorsal angulation of lunate resulted. G, Aninlay (l~’.usse) bone graft with internal fixation resulted union and good functional result.

were considered to be unsuccessful results. The overalltime for union averaged 4.4 months in this series, butunion depended on the grafting technique and the typeof fracture.

Complications from bone grafting of scaphoid non-unions were avascular necrosis (four patients) arthrosiswith union (14 patients) and nonunion (22 patients)(Table II). Of the persistent nonunions after treatment(22 fractures), four were associated with avascular ne-crosis and 16 with secondary radioscaphoid arthritis.Of the nine peg graft failures, five required reoperation(arthrodesis in two, silicone rubber scaphoid in one,and volar inlay graft in two). Four of the six nonunionsafter w~lar inlay graft were regrafted (Fig. 5), and tworemained untreated. One dorsal inlay graft was re-grafte~l successfully by the same technique. Two screwosteosynthesis faiiures were treated by a dorsal inlaygraft: and union was obtained.

Avascular necrosis identified at the initial treatment

348 Cooney et al.

The JournaHAND SURGEI;

Radial

incision E~xtensor Pollicis Longus

Extensor Potlicis

pno~a \ ~Snuff Box

Styloid processof radius

Nonunion site

Fig. 6. Matti’s original dorsal inlay approach te. scaphoid nonunion was modified by Dobyns andLipscomb. The technique consists of: (A) a radial approach between the first and second dorsalretinacular compartments; (B) corticocanceIlous bone graft (20 by 5 mm) from the radial styloid; radial styloidectomy, preparation, and excavation of a dorsoradial trough in proximal and distalfragments with removal of all avascular bone; ~md (D) inlay of the corticocancellous bone graft,which is countersunk to prevent displacement; internal fixation of the fracture fragments. E, A caseillustration of an established nonunion through the mid distal third junction of the scaphoid. F,Treatment by radial styloidectomy, radial inlay bone graft with Kirschner wire fixation of thefracture fragments and radiocarpal joint. G, Fracture union without pain or arthrosis was present 2years later.

5. No. 4~980 Nonunion of scaphoid 349

Fig. 7. A and B, Early result (6 months) after compressio:a screw fixation of scaphoid nonunion.Bone grafting was not performed. C, At 18 months screw migration had occurred and there was noevidence of scaphoid union. D, A volar inlay graft was performed, and 3 months later (21 monthsfrom the original injury) progression toward scaphoid union was present.

associated with scaphoid nonunion in 13 fractures.union was obtained by bone grafting in 11 of

13. In the two patients with failed union, the frac-i.ture Was a displaced, proximal pole fracture--one wastreated by the peg graft technique and the other was

tile volar inlay technique. Four other patientsavascular necrosis after operation. Union

occurred in two (both dorsal inlay grafts), with eventualof the proximal fragment. The two

developed avascular necrosis and nonunion afterby the dorsal peg technique. Both involved

proximal third of the scaphoid, and one also was

Internal fixation was used in 21 of 44 volar inlaywhen motion was present at the fracture site

instability demonstrated). Union resulted inof these fractures. In the dorsal peg technique,

fixation was performed in only four of the 18fractures. Ten of the 18 fractures were displaced non-

unions. The four that had pin fixation united. The re-maining displaced fractures (six) without pin fi~ation

accom~ted for five of the nine nonunions. All of thefractures treated by the dorsal inlay technique had twoor more Kirschner wires, and 20 of the 22 united.

Artlarosis despite fracture union was present to a mild--to moderate degree in 14 patients (16%). Occasionalpain usually associated with heavy labor was present infour patients (10%) with a united scaphoid nonuniontreated by the volar inlay technique, four patients(20%) with united dorsal inlay graft, and five of ninepatients with united dorsal peg graft. Roentgenogramsrevealed that 9% of the wrists with volar inlay graftsand 14% of the dorsal inlay grafted had radioscaphoidarthrosis. Nearly half (44%) of those treated by thedorsal peg method had roentgenographic evidence ofradioscaphoid arthrosis.

Grip strength for the entire group was 36 kg afteroperation--from 27 kg before operation for an increaseof 30% in comparison with the uninjured side (average,45 kg). Wrist motion (preoperative versus postopera-tive) changed little. Dorsiflexion was unchanged, andvolar flexibn lost 5° of motion. The volar and dorsal

350 Cooney et al.

inlay grafts resulted in more improvement in gripstrength than did the dorsal peg technique, and thisappeared to be related to the roentgenographic appear-ance of arthrosis and to the higher incidence of fracturenonunion.

Discussion

Bone grafting to restore stability and union is theprocedure of choice for established scaphoid non-unions.a, r, Exceptions to this are generalized and se-

vere posttraumatic arthritis, a small and avascular prox-imal fragment, and the patient who is asymptomatic orsedentary.la, ~6, i7 The natural history of scaphoidnonunion is that late morbidity from pain, loss of mo-tion, and grip weakness usually will result if the symp-tomatic patients are not treated.’-’, s, ~e. 16

When bone grafting is elected for established non-union, the results from this series indicate that a dorsalor volar inlay graft is best. This technique of inlay bonegrafting was originally described by Matti9 in 1932. Herecommended a dorsal approach. By this technique,direct inspection of the fracture nonunion was per-formed; a trough was curetted in the distal and proximalfragments, removing sclerotic bone; and a cancellousiliac crest bone graft was inserted. Russe~a modifiedthis method by recommending a volar approach. One ofus (J.H.D.) used the dorsal inlay technique by modify-ing the radial styloidectomy approach of Barnard andStubbins and using the radial styloid as the source ofcorticocancellous inlay bone graft (as described byMatti) (Fig. 6).

This technique consists of exposing the radial styloidby dissecting off the abductor pollicis longus and ex-tensor pollicis brevis with the first extensor compart-ment. The radial sensory nerve and the radial arterywith its branches to the scaphoid are carefully pro-tected. A sagittal saw is utilized to remove a cor-ticocancellous bone graft (approximately 20 by 5 by mm). The stylold itself is removed using an oblique cutcalculated to enter the radial articular surface ulnar tothe position of the scaphoid fracture. With the fracturesite exposed, an appropriate-size trench is cut into thedorsoradial or radial aspect and deepened by curettageof the distal and proximal fragments. Microscopicexamination can help to determine if there is avascularbone. The bone graft, along with cancellous bonechips, is then packed into the trough, and each endis countersunk. Correct reduction of the fracture ischecked, and any malunion is reduced. MultipleKirschner wires are inserted to stabilize first the site ofthe nonunion and then the radiocarpal joint. With un-

The JourrHAND SURGI

stable nonunions it is best to realign the dorlunate, pin the proximal scaphoid fragment to theduced lunate, and reduce the distal scaphoid fragme~onto the fixed and well-aligned proximal scaphoidIt is important that care should be taken that the frac~fragments are not distracted dorsally and that radio2~’graphic malalignment is not present. For corerfractures, bone grafting may be necessary to cornsate for bone loss in order to achieve an accurate rection of the malunion. The volar (Russe) approachperformed similarly, with the exception that radstyloidectomy is not necessary and internal fixationless frequently indicated because the volar bone graftwill close the dorsal cortex angulation and provide in.:herent stability. A large iliac corticocancellous bonegraft is preferred for this approach, but care musttaken to maintain anatomic alignment. ,

These methods were superior, in our hands, to thepeg graft technique credited to Murray~° in 1934. Thmethod had a 50% incidence of nonunion (nine of 18fractures). It was performed by a number of surgeonsfor both displaced and undisplaced fractures. A hiincidence of radioscaphoid arthrosis (44%) occurredafter this method than after the volar (9%) and dorsalinlay (14%) techniques. Murray~r reported that 96% ofhis patients had union, but stressed a largegraft, accurate plac.ement through a substantial drillhole, and fracture immobilization until roentgeno-graphic union. In this series immobilization withwas used in only four of 18 patients, and care wasalways taken to differentiate the stable from theble group. Others have reported reduced rates of unionsimilar to our experience with this technique,and the differences in these series from Mrfi’ay’s re.ports cannot be well explained.~7 .

Compression screw osteosynthesis has beenmended by several authors for scaphoid nonunion.8~.McLaughlin and Parkes16 used this method in fivetients with pseudarthrosis and found that noneunion. (They preferred the Russe volar inlaynique.) Gasser’s series of ASIF screw fixationsisted of 20 patients. Two had acute fractures,had fractures that were less than 6 months old. Allcandidates, in our opinion, for castalone. Of the 20 patients with scaphoid screw fixatiori;he had 11 patients with complete union and ninetients with nonunion. Seven of thosethe group that was less than 6 months from theinjury. In the present series there was onlywith union of the six with nonunions treated bycompression screw. Four of the six patients with

No..41980

Nonunion of scaphoid 351

Fig. $. Fracture of scaphoid proximal pole with nonunion. A and B~ Established nonunion ofproximal one-third with displacement evident on the lateral view. C, A tomogram showed viablefragment, and bone grafting was recommended. D and E, A volar Russ~ approach was performed,with curettement of the entire proximal pole and trough formed for placeCaent of a corticocancellousiliac pone graft. Note the preserved laterat-volar artery (arrow). F and G, Union was observed at 4m~rtt9-~ ~ t~ra~grams and was confirroed at the 5-year follow-up.

352 Cooney et al.The Journal

HAND SURGE~

Fig. 9. Avascular necrosis of the proximal portion of the scaphoid and nonunion (A). After dorsalinlay bone grafting with internal fixation (B), union resulted (C), and resolution of avascularsegment progressed (D) with fracture healing and revascularization.

tures had evidence of retrograde screw migration (Fig.7), and two had undergone reoperation. Thus late in-stability without bone healing may be anticipated fromthis technique in patients who have established non-union. None of these nonunions had a substantial bonegraft (matchs~ick only), and, if a larger corticocancel-lous graft could be added to the compression screwfixation, the union rate might be expected to increase.

The location of the fracture3’ 19 and classification ofstable and unstable nonunions are important factors inthe prognosis and the selection of the surgical tech-nique.4 The union rate was 64% for proximal pole frac-tures, 80% for mid third fractures, and 100% for distalthird fractures. The two patients with avascular necro-sis at follow-up had fractures of the proximal pole. Weprefer the volar approach for most proximal pole frac-tures (Fig. 8), because the inlay graft can be insertedmore easily than by a dorsal approach, which wouldrequire a generous radial styloidectomy for exposure.

Union was more frequent as a result of ......approach than from either the dorsal inlay or thegraft approach when the fracture was in the proximapole.

Displaced scaphoid fractuyes had a union rate ofin comparison with the rate of 85% for undisstable fracture nonunions. McLaughlin and Parkes,1.

Verdan and Narakas,3 Fisk,2° and others2’ 21, 22emphasized the need to classify scaphoidinto displaced and undisplaced and the importancethe lunate angulation that accompaniesfractureplacement. More recently, Dobyns andhave brought attention to the dorsiflexion collapseformity that results from displaced scaphoid fracture,’Eddeland and associates1~ noted the greaterof pseudarthrosis when displacement exceeded 1and this was true in our series2 The importance of thelateral roentgen0gram to assess carpal instability cannot ibe overstressed. Dorsal tilt or rotation of the lunate

No. 4

1980

when the scaphoid is displaced. If uncertaintyroentgenographic views such as distraction in

traps, radioulnar deviation views, or firm corn-gripping anteroposterior and lateral views mayadditional information in determining the de-

of displacement and instability. Our experiencethat, when displacement is present, a volar or

: combined with internal fixation of not onlyscaphoid but also the lunate in its reduced position

distal radius increases the rate of union and re-the risk of secondary arthrosis. When radio-

arthrosis is evident, we prefer the dorsal ap-with radial styloidectomy, direct inspection and

ti(-,~ of the fracture site, a dorsal inlay bone graft,intcl-nal fixation of the fracture and the carpus.

role of fracture stabilization by internal fixation.ins should be emphasized, particularly in the dis-

fractures. In the displaced nonunion group,were fewer complications and a higher rate ofin the Russe volar graft with pins and the dorsalgraft with pin techniques than in those fractures

without internal fixation. Cast support alonenot appear to provide sufficient immobilization.lc~z~th of time in cast support for displaced frac-.aith pins or screw fixation included should be a

of 6 weeks, and longer (up to 3 months) forpole fractures. We prefer a long-arm cast thatthe thumb for 6 to 8 weeks, and then a short-

thumb spica cast until there is roentgenographicof union. The concept that roentgenographic

after clinical union is not supported by the)hic results in many of these studies. The sin-amportant factor in the failure of bone grafting

in discontinuing the cast support.2" 19, 24supply to the scaphoid has been well de-

Taleisnik and Kelly.2~ We studied the effectapproach on the blood supply in four

hands. Since both the lateral volar and dorsalbranches from the radial artery to the scaphoid

in size and cross-anastomose in supplyingtwo-thirds of this bone,25, 26 it was con-

that neither the dorsal nor the volar approachto the other from a vascular standpoint.

necrosis, which was present in 13 patientsoperation, was resolved in 11--six by the volar

the dorsal approach. The two patients whoto obtain union and who had persistent avas-

necrosis had small proximal fragments; one of theialso had displaced fragments. Russe13 and Verdan

~a have noted that avascular necrosis re-

fracture union, and this was also tru6 in our(Fig. 9). If the surgical approach decreased vas-

Nonunion of scaphoid 353

cularity to the scaphoid, we would have expected agreater incidence of avascular necrosis and nonunion.

Concl:usions

This study of the bone-grafting treatment for scaphoidnonun!ions revealed that long-term results from thebone grafting of scaphoid nonunions were favorable byboth dorsal and volar inlay grafts, but that peg graftshad a higher rate of nonunion and secondary arthritis.Grip strength was improved (inlay grafts greater thanpeg grafts), but wrist motion was not changed. Scaph-oid fractures should be classified into stable or unstableon the basis of roentgenographic evidence of displace-ment of greater than 1 mm or of instability collapsepatterns (dorsal rotation of the lunate) or both. Internalfixation of the unstable nonunion is recommended, andthe carpus should be included with reduction of thelunate. Compression screw osteosynthesis providesstability but not union; long-term results are discourag-ing. The surgical approach did not significantly influ-ence fracture union, and avascular necrosis was rareafter these procedures. Nonunions should be immobil-ized until there is roentgenographic evidence of union,and a minimum of 4 months is recommended for stablenonuniions.

REFERENCES

1. Cobey MC, White RK: An operation for non-union offractures of the carpal navicular. J Bone Joint Surg28:757-64, 1946

2. Mazet R Jr, Hohl M: Fractures of the carpal navicular:analysis of ninety-one cases and review of the literature.J Bone Joint Surg [Am] 45:82-112, 1963

3. Verdan C, Narakas A: Fractures and pseudarthrosis ofthe: scaphoid. Surg Clin North Am 48:1083-95, 1968

4. Stewart MJ: Fractures of the carpal navicular (scaphoid):a report of 436 cases. J Bone Joint Surg [Am] 36:998-1006, 1954

5. Adams JD, Leonard RD: Fracture of the carpal scaphoid:a l:tew method of treatment with a report of one case. NEngl J Med 198:401-4, 1928

6. Bamard L, Stubbins SG: Styloidectomy of the radius inthe, surgical treatment of non-union of the carpal navicu-lar: a preliminary report. J Bone Joint Surg [Am] 30:98-10:2, 1948

7. Bentzon PGK, Randlov-Madsen A: On fracture of thecarpal scaphoid. Acta Orthop Scand 16:30-9, 1946

8. Gasser H: Delayed union and pseudarthrosis of the carpalnavicular: treatment by compression-screw osteosyn-thesis. J Bone Joint Surg [Am] 47:249-66, 1965

9. Matti H: Technik und Resultate meiner Pseudarthrosen-operation. Z Chir 63:1442-53, 1936

10. Murray G: Bone-graft for non-union of the carpalscaphoid. Br J Surg 22:63-8, 1934

354 Cooney et al.

11. T6rngren S, Sandqvist S: Pseudarthrosis in the scaphoidbone treated by grafting with autogenous bone-peg. ActaOrthop Scand 45:82-8, 1974

12. Eddeland A, Eiken O, Hellgren E, et al: Fractures of thescaphoid. Scand J Plast Reconstr Surg 9:234-9, 1975

13. Russe O: Fracture of the carpal navicular: diagnosis,non-operative treatment, and operative treatment. J BoneJoint Surg [Am] 42:759-68, 1960

14. Lipscomb PR, Dobyns JH: Personal communication15. Maudsley RH, Chen SC: Screw fixation in the manage-

ment of the fractured carpal scaphoid. J Bone Joint Surg[Br] 54:432-41, 1972

16. McLaughlin HL, Parkes JC: Fracture of the carpalnavicular (scaphoid) bone: gradations in therapy basedupon pathology. J Trauma 9:311-8, 1969

17. Murray G: End results of bone-grafting for non-union ofthe carpal navicular. J Bone Joint Surg 28:749-55, 1946

18. Mazet R Jr, Hohl M: Radial styloidectomy and styloidec-tomy plus bone graft in the treatment of old ununitedcarpal scaphoid fractures. Ann Surg 152:296-302, 1960

19. Unger HS, Stryker WC: Nonunion of the carpal navicu-

The Journal iHAND SUR

lar: analysis of 42 cases treated by the RusseSouth Med J 62:620-2, 1969

20. Fisk GF: Carpal instability and the fractured scaAnn R Coll Surg Engl 46:63-76, 1970

21. Dickison JC, Shannon JG: Fractures of thescaphoid in the Canadian army: a review and commetary. Surg Gynecol Obstet 79:225-39, 1944

22. Gilford WW, Bolton RH, Lambrinudi C: The me~nism of the wrist joint: with special reference to fracturof the scaphoid. Guys Hosp Rep 92:52-9, 1943 ¯

23. Dobyns JH, Linscheid RL: Fractures and dislocationsthe wrist, in Rockwood CA Jr, Green DP, editors:tures, vol 1. Philadelphia, 1975, JB Lippincott Co,345-440

24. Watson-Jones R: Fractures and joint injuries, vol 2. ed.~Baltimore. t955, Williams & Wilkins Co

25. Taleisnik J, Kelly PJ: The extraosseous and intraosseousblood supply of the scaphoid bone. J Bone Joint Surg[Am] 48:1125-37, 1966

26. Obletz BE, Halbstein BM: Non-union of fractures of~carpal navicular. J Bone Joint Surg 20:424-8, 1938

INFORMATION FOR AUTHORS

Most of the provisions of the Copyright .Act of 1976 became effective on January 1,1978. Therefore, all transmittal letters must be accompanied by the following statement,signed by each author: "The undersigned author(s) transfers all copyright ownership the manuscript entitled (title of article) to the American Society for Surgery of the Hand

in the event the work is published. The author(s).warrants that.the article is original, not under consideration by another journal, and has not been previously published."Authors will be consulted, when possible, regarding republication of their material.