- Email: [email protected]
Treatment of midshaft clavicular nonunion with plate fixation and autologous bone grafting
Treatment of midshaft clavicular nonunion with plate fixation and autologous bone grafting Bo Sanderhoff Olsen, MD, Michael T. Vaesel, MD, and Jens O. Sojbjerg, MD, Aarhus, Denmark
We studied the results of 16 consecutive midshaft clavicular nonunions operated on at the Shoulder and Elbow Clinic during the period from 1990 to 1993. All patients were treated with rigid 3.5 mm plate fixation and autologous cancellous bone grafting. Union of the fractures was achieved in all except one case, with a reconstruction ratio (restoration of bone length) of 0.96 (range 0.88 to 1.03). At follow-up 12 of 16 patients had returned to their preinjury activity level and according to the Constant score had obtained an excellent result. Two patients were graded as good, one as fair, and one had a failure. Thirteen of 16 patients were satisfied with the cosmetic outcome, assessing their cosmetic result as either good or excellent. Rigid plate fixation and restoration of clavicular length with autologous cancellous bone graft is recommended for the treatment of symptomatic clavicular midshaft nonunions. (J SHOULDERELBOWSURG 1995;4:337-44.)
Fracture of the clavicle is common, comprising between 5% and 16% of all fractures seen in a casualty department.* The incidence of nonunions is reported to be between O. 1% and 5%. 2' 6-8, 12, 15-17 Most nonunions and fractures appear in the middle one third of the clavicle, s' z. 15 A clavicular nonunion is usually diagnosed by lack of fracture healing 4 to 6 months after injury.S, 12, 17, 19, 20 It is generally agreed that surgery is indicated only in symptomatic nonunions.'~ ,6, 19, 2o Previous reports describe moderate to severe pain, grating, cosmetic deformity, and weakness of the shoulder in most of the symptomatic cases. 5" 8, lO, 12, 15, 18-2oVarious surgical and nonsurgical techniques have been described for the treatment of this potentially disabling condition.2, ~-8, lO, 12, 17-2oExternal or internal fixation of a clavicular nonunion is demanding because *References 2, 7, 8, 12, 15, 16, 18. From the Shoutder and Elbow Clinic, Department of Orthopaedic Surgery, Universib, Hospital of Aarhus. Reprint requests: Bo Sanderhoff Olsen, MD, Orthopaedic Research Fellow, Biomechanics Laboratory and Shoulder and Elbow Clinic, Universib, Hospital of Aarhus, Randersvej 1, DK-8200 Aarhus N, Denmark. Copyright @ 1995 by Journal of Shoulder and Elbow Surgery Board of Trustees. 1058-2746/95/$5.00 + 0 3 2 1 1 / 6 7 6 6 9
of the basic anatomy and biomechanics of the bone. s' 1o During recent years several authors have reported good results after surgical treatment of clavicular nonunions by using plate fixation and autologous iliac crest bone grafting.2, s, 6, 8-,3, 17, 18, 2o However, most studies include only a limited number of cases treated over a period of several years. 2" 5, 6, 8-1s, lZ, 18, 2o This treatment is claimed to be successful with a high ratio of fracture healing, but only two reports evaluate the functional and cosmetic outcome of the surgery. 1~ 12 In this retrospective study we reviewed the functional and cosmetic outcome of 16 consecutive patients with symptomatic clavicular midshaft nonunions treated with rigid internal fixation and autologous cancellous bone grafting.
PATIENTS AND METHODS Seventeen consecutive patients with a clavicular midshafl nonunion were treated at the Shoulder and Elbow Clinic, Aarhus University Hospital, during the period from early 1990 to late 1993. The patient data are presented in Table I. One patient was lost to follow-up and was excluded from the study, leaving 16 patients for study. Ten patients were men, and six were women. Twelve fractures were right sided and four left sided. The dominant 337
338
J. Shoulder Elbow Surg. September/October 1995
Olsen, Vaesel, and Sojbjerg
Table I Data for all patients in study Patient No.
Age at operation (yr)/sex
Injury/dominant side
Mechanism of injury
Duration of nonunion (wk)
Initial treatment
1
35/M
R/R
Car accident
2
35/F
R/R
Sport/fall
3
30/M
R/R
Bicycle accident
20
Figure of eight 14day delay
4
18/F
L/R
Car accident
18
Plating
5
49/M
R/R
Motorcycle accident
33
Arm sling 3 days
6
55/F
R/R
Bicycle accident
44
Figure of eight < 14 days
7
49/M
R/R
Bicycle accident
60
Figure of eight
8
32/M
R/R
Car accident
9
25/M
R/L
Bicycle accident
26
Figure of eight
10
38/F
L/R
Bicycle accident
25
Figure of eight 3-day delay
11
48/M
L/R
Motorcycle accident
22
Figure of eight
12
19/F
R/R
Car accident
50
!3
31/M
R/R
Motorcycle accident
90
Arm sling 14-day delay Arm sling 10 wk
14
43/M
R/R
Simple fall
65
Arm sling <14 days
15
13/F
L/R
Simple fall
164
Figure of eight 7 wk
16
35/M
R/R
Car accident
988
Figure of eight
16
Armsling <1 wk
144
Figure of eight
390
Arm sling < 14 days
Excellent,Excellent result, Radiographic union; z~CS(Constant scoring) <10. Activity normalized; Good, Good result: Radiographic union. ACS <20. AActivity _<2;Fair,Fair result: Radiographic union. ACS <30. z~J~ctivity_<4;Failed,Failed result: No radiographic union.
arm was affected in 11 patients. In three patients the fracture was sustained after a simple fall, in five it was the result of a bicycle accident, and in nine it was caused by a high-energy traffic accident. All fractures were classified according to Allman 1 and Neer TM as belonging to group 1 (midshaft) clavicular fractures. All fractures were initially treated conservatively with a figure-of-eight bandage or an arm sling for 3 to 61 days, except for one fracture primarily treated elsewhere with a compression plate. The time from trauma to final surgery was a mean of 135 weeks (range 16 to 988 weeks). The mean age of the pa-
tients at surgery was 34 years (range 13 to 55 years). Nine patients had an atrophic nonunion and six a hypertrophic nonunion. One patient had a failed osteosynthesis. On gross examination nine of the nonunions were displaced, presenting an overlap, diastasis, or both equal to or greater than the diameter of the clavicle (Fig. 1, A). Four nonunions presented with angulation exceeding 25 ~ The indication for surgery in 14 of 16 patients was pain in the affected shoulder during exercise or at rest. Seven patients had moderate pain only present at exercise, and seven described the pain as severe with pain at rest. Two patients were free
J. Shoulder Elbow Surg. Volume 4, Number 5
Fracture type and overlap
Olsen, Vaesel, and Scsjbjerg
Length of Postoperative clavicular Outcome follow-up (mo) length ratio of operation
Hypertrophic
51
0.97
Excellent
Hypertrophic > 1 bd overlap Hypertrophic Angulation 30 ~ Failed plating
30
0.92
Excellent
30
0.98
Excellent
34
0.93
Fair
Atrophic >2 bd diastasis Atrophic >3 bd overlap
30
0.92
Excellent
39
0.96
Good
Atrophic >2 bd overlap + diastasis Atrophic
21
0.99
Excellent
24
1.01
Excellent
Atrophic >2 bd diastasis Hypertrophic Angulation 30 ~ Atrophic >4 bd overlap Atrophic >4 bd overlap Atrophic >3 bd overlap Hypertrophic Angulation 25 ~ Hypertrophic Angulation 60 ~
18
1.00
Excellent
29
0.96
Excellent
28
1.03
Excellent
12
1.00
Fair
12
0.98
Excellent
12
0.96
Excellent
12
0.88
Excelbnt
14
0.90
Failed
Atrophic >3 bd overlap and diastasis
of pain but had a grating sensation during heavy work. In these two patients the primary indication for surgery was cosmetic. At operation the patient was placed in the semiseated position with the head turned to the opposite shoulder. The anterior iliac crest was prepared to obtain the bone graft. All patients received prophylactic antibiotics. A horizontal skin incision along the caudal edge of the clavicle was used. The normal clavicle and as much as possible of the nonunion was~ exposed subperiosteally. In the hypertrophic nonunions the hypertrophic callus was reduced to normal size, and interposed tissue was excised. In the atrophic cases the sclerotic ends were resected, and in both types the medullary canal was opened and reamed with a drill.
339
Comments Normal pain free ROM Normal pain free ROM Normal pain free ROM Normal pain free ROM. Sequelae from pelvic fracture Normal pain free ROM Pain in internal and external rotation Decreasedinternal rotation Pain free ROM Normal pain free ROM Normal pain free ROM Normal pain free ROM Normal pain free ROM Abduction 150 ~ Pain during ROM Normal pain free ROM Normal pain free ROM Prior ipsilateral fracture Pain free normal ROM Painful decreased ROM
After the nonunion was released, the shoulder tended to slide posteriorly, thereby restoring the length of the clavicle. A 3.5 mm plate was contoured to shape the clavicle exactly and was temporarily fixed to the clavicle by two bone clamps. The distance from the acromial tip to the midsterhum was palpated bilaterally to approximate the correct length of the clavicle, and the plate was secured to the bone with a minimum of six screws (Fig. 2). Autologous cancellous bone was harvested from the anterior iliac crest. The bone graft was packed into the clavicular defect, and in two cases with cortical grafting the intercalcary segment was fixed with a screw. The trapezius and deltoid fascia were carefully repaired. An arm sling was used for postoperative immo-
340
J. Shoulder Elbow Surg. September~October 1995
Olsen, Vaesel, and S~jbjerg
Table II Cosmetic score Cosmetic outcome
Preoperative
Postoperative
Excellent Good Acceptable Poor Bad
0 0 1 7 8
3 10 1 1 1
bilization for at least 21 days. Only passive motion of the shoulder not exceeding 90 ~ of elevation was allowed for the first 6 weeks. Heavy work was allowed after 3 months, and in general contact sports were allowed after 6 months. In 13 patients removal of the osteosynthesis material was necessary for pain or cosmetic reasons at an average of 11.6 months (range 3.5 to 27 months) after the operation. The follow-up time was a mean of 24 months (range 12 to 51 months). At follow-up, x-ray films, including those of the complete shoulder girdle, which allowed comparison of the clavicular length on the operated and nonoperated side, were obtained. This x-ray examination enabled evaluation of the restoration of clavicular length, which was calculated as the ratio between clavicular length on the injured and the noninjured side and the exact difference in millimeters. The cosmetic results were assessed by the patients and graded on a 1 to 5 scale, 1 being excellent and 5 being bad (Table II). In all the patients the activity scores were recorded for the time before the injury, after the trauma but before the operation, and after the operation. Patients were graded on a 1 to 8 scale, 1 expressing the highest activity level and 8 the lowest (Table III). Finally, the Constant shoulder score for assessing shoulder function was applied. 3' 4 This scoring allows evaluation of differences between injured and noninjured shoulders, recording pain, range of movement, and the objective force during combined abduction and elevation; these factors are measured with the Isobex traction device (Isobex, Curmed, Zurich Switzerland) (Table IV). The results were evaluated with a Student t test, paired and unpaired (SOLO; BMDP statistical software, Los Angeles, Calif.) and were considered significant at p < 0.05. RESULTS Solid healing was achieved in 15 of 16 patients (Figure 1, C). After the operation a minor clavicu-
lar shortening on the injured side was observed after x-ray evaluation. An average shortening of 6 mm (range -4 to 16 mm) was measured, with a mean clavicular reconstruction ratio of 0.96 (range 0.88 to 1.03) (Table I). Before the operation several patients had a marked shortening of their shoulder girdle as a result of the nonunion, with an overlap of the fracture ends of up to four times the bone diameter of the clavicle. This approximates a shortening of up to 80 mm (Table I and Figure 1, A, B). All the patients had forward and downward dropping of the injured shoulder compared with the noninjured side before the operation. This condition was usually described as hanging and shortening of the shoulder. Some patients also observed the medial fragment pointing proximally, threatening the integrity of the skin. Surgery cured the reported dropping in 15 of 16 patients, but in two of these patients a hump was observed at the former fracture site, reducing the cosmetic outcome. One patient (case 16) had an unchanged cosmetic score (Table II). After the operation a median gain of two classes on the cosmetic scale was reported compared with the preoperative observations (p < 0.001, paired t test). Thirteen of the 16 patients were satisfied with the cosmetic outcome of surgery, grading their result as either good or excellent (Table II). Fourteen of 16 patients had pain and functional disability before operation. After the operation 11 patients were free of pain, three had a mild grating sensation in the shoulder during exercise, and two patients had moderate pain, with pain being present only during heavy work. No cases of severe or disabling pain were seen after the ope r a t i o n . Twelve of 16 patients returned to their pretraumatic activity level after surgery (Table III); these were all rated as having an excellent surgical outcome (Table I). Three of the remaining four patients (cases 4, 6, 12) had some improvement but at the time of follow-up had not been able to resume their pretraumatic activity level, although healing of the clavicular nonunion occurred in all these patients. In one patient (case 16) no union was achieved. This patient had the longest period from trauma to final surgery (19 years). He had previously been treated with resection of a part of the medial fragment. A cortical graft from the iliac crest was used during the plating procedure to span the gap between the bone ends. Healing of this graft was
J. Shoulder Elbow Surg. Volume 4, Number 5
Olsen, Veesel, and S~jbjerg
341
Figure 1 Radiograph of patient 6. A, Preoperativex-ray film showing large overlap of atrophic nonunion. B, Postoperativex-ray film. C, Postoperativex-ray film showing operative outcome after removal of plate. never achieved. The compression plate had to be removed because of loosening and protrusion of the screws, threatening the skin. At follow-up this patient had a reduced and painful range of mo-
tion. He is now claiming workman's compensation. A median gain of three classes in the activity scale was reported after the operation compared
342
J. Shoulder Elbow Surg. September~October 1995
Olsen, Vaesel, and S~jbjerg
j Figure 2 Reconstruction of clavicular nonunion with plate and autologous
cancellous bone graft in gap.
Table III Activity score Activity level 1. 2. 3. 4. 5. 6. 7. 8.
ProFessionalsport (shoulder-demanding) NonproFessional sport (shoulder-demanding) Professionalsport (non-shoulder-demanding) Shoulder demanding profession Nonprofessional sport (non-shoulder-demanding) Non-shoulder-demanding profession Unemployed, not because of shou(der Unemployed, because of shoulder
with the posttraumatic activity level (p < 0.001, paired ttest). Evaluation of the pretraumatic activity compared with the postoperative activity showed a median decrease of one class (p < 0.05, paired t test). The Constant score of the injured and the noninjured shoulders at follow-up are listed in Table IV. These scores show a significantly reduced score for the injured shoulder compared with that of the noninjured shoulder (p < 0.02, paired t test). Patients operated later than 40 weeks after the initial injury had a significantly poorer outcome (p < 0.02, t test) than the patients operated in an earlier phase. Furthermore the atrophic nonunions had a poorer outcome compared with the hypertrophic nonunions (p < 0.03, t test). Apart from this finding we found no correlation between the outcome of surgery evaluated through the Constant score compared with age, sex, preoperative or postoperative clavicular shortening, and dominant or
Pretraumatic
Posttraumatic
Postoperative
0 6 0 8 2 0 0 0
0 1 0 0 2 5 0 8
0 3 0 8 4 1 0 0
nondominant arm. No neurologic dysfunction or thoracic outlet signs were observed after the operation. In this study no cases of superficial or deep infections were observed. Except for patient 16, who had loosening of the screws, no complications developed. DISCUSSION The clavicle provides stability and force to the shoulder joint by serving as a bony stabilizer and as a link between the thorax and the shoulder girdle, s This explains the described functional and cosmetic consequences of nonunion after a clavicular fracture, which were also observed in this study, lO, 15, 19 This study reports good functional outcome of plate fixation and autologous cancellous bone grafting in symptomatic, clavicular, midshafl nonunions, with grading of the functional outcome as excellent in 12 of 16 patients, and only one failure. This result is in accordance with previous
J. Shoulder Elbow Surg. Volume 4, Number 5
reports. 8, lO, 12 Furthermore this study exhibits good or excellent cosmetic results in 13 of 16 patients, with only two patients being dissatisfied with the final cosmetic outcome. In this study all but one patient returned to their pretraumatic occupations. This result corresponds to that in the report from one other author. 8 Several factors have been reported predisposing to nonunion after a clavicular fracture: (1) inadequate initial immobilization after closed or open reduction, s' lo, 13, 19 (2) severity of fracture displacement/' 8, lo, is, 19, 2o (3) primary open reduction of the fracture, s' lO, is, 19 (4) age (children rarely present with posttraumatic clavicular nonunions).S In this series 8 of 16 patients were noncompliant with the initial immobilization of 21 days. Some of these patients had an initial delay before immobilization of several days; others immobilized the shoulder girdle only during a limited period of less than 14 days (Table I). Only 7 of 16 cases underwent initial immobilization for at least 21 days. This finding supports the theory that inadequate immobilization of clavicular fractures favors the development of nonunions, s" lo, 13, 14 CraigS stated that fractures of the clavicle should be immobilized until healing, reported as being 4 to 6 weeks in young adults and 6 to 8 weeks in older patients. Nine of 16 patients initially had severe grades of fracture displacement presenting as overlap of more than one times the bone diameter. Furthermore four patients had angulation exceeding 25 ~. This finding confirms reports stating that the degree of displacement is an important predictor of the development of nonunion after a clavicular fracture.S, lo, is, 2o This study included only one child who was 13 years old. When this rate is compared with the rather high incidence of clavicular fractures among children, it tends to support the theory that children rarely have clavicular nonunions, s In this study the Constant score was significantly lower in the affected shoulder compared with that in the nonaffected shoulder, although the dominant arm was more frequently affected than the nondominant arm. Other authors found no association between the injured side and hand dominance of the patient. 1~ 2o A long period of conservative treatment of a symptomatic clavicular nonunion is recommended by some authors, who state that atrophic nonunions become less symptomatic with time and
Olsen, Vaesel, and Sojbjerg
343
Table IV Constant score at follow-up Patient No.
Injured
Noninjured
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
97 88 99 81 100 73 87 92 100 89 90 72 97 98 79 65
98 85 100 88 100 88 89 100 100 90 97 100 100 100 82 100
therefore do not necessarily require surgery. 2~ This belief was not supported by this study, because all atrophic nonunions were observed to be symptomatic in spite of a long follow-up. Furthermore the two patients with only cosmetic symptoms had hypertrophic nonunions.12 In this study the atrophic nonunions had a significantly poorer outcome than did the hypertrophic cases when compared according to Constant score and activity level. The only failure in this study was initially an atrophic nonunion. These results support the opinion of Connolly and Dehne 2 that the atrophic nonunions have a poor prognosis. This study also indicates that early operation of clavicular nonunions gives a better functional outcome.
In two cases free cortical lilac grafts were used to span clavicular defects. One of these cases failed. Other authors using free tibial or lilac grafts reported failure rates of 50%. is Different authors have also reported subjective weakness in abduction of the affected shoulder. Whether the weakness is caused by pain or by clavicular shortening is not specified. 8' 14, 2o No significant objective differences in force transmission over the affected shoulder girdle compared with that over the normal side were observed in this study. Force transmission was measured as part of the Constant scorings. The patients in this study all underwent careful reconstruction. This reconstruction yielded small differences in length between the operated and normal sides (Table I). On the basis of this series we conclude that the
344
J. Shoulder Elbow Surg.
Olsen, Vaesel, and S~jbjerg
September~October
restoration of clavicular length is essential to obtain good functional and cosmetic results. This study indicates that for a clavicular midshaft nonunion, the present operative technique offers a safe surgical method. This method has a predictably good end result concerning healing of the nonunion, which was achieved in 94% of the cases, and a good functional and cosmetic outcome for the patient. The results from this study and earlier reports suggest that rigid plating with autologous cancellous bone grafting is the treatment of choice for the patient with a symptomatic clavicular midshafl nonunion.* REFERENCES
7. 8. 9. 10. 11. 12. 13.
1. AIIman FLJ. Fractures and ligamentous injuries of the clavicle and its articulation. J Bone Joint Surg Am 1967;49A:774-84. 2. Connolly JF, Dehne R. Nonunion of the clavicle and thoracic outlet syndrome. J Trauma 1989;29:112732. 3. Constant CR. Assessment of shoulder function. Orthopade 1991 ;20:289-94. 4. Constant CR, Murley AH. A clinical method of functional assessment of the shoulder. Clin Orthop 1987; 160-4. 5. Craig ED. Fractures o~ the clavicle. In: Rockwood CA, Matsen FA, editors. The shoulder. Philadelphia: Saunders, 1990:367-412. 6. Edvardsen P, Odegard O. Treatment oF posttraumatic *References 6, 8, 10, 12, 13, 18, 20.
14. 15. 16. 17. 18. 19. 20.
1995
clavicular pseudarthrosis. Acta Orthop Scand 1977;48: 456-7. Eskola A, Vainionpaa S, Myllynen P, Patiala H, Rokkanen P. Outcome of clavicular fracture in 89 patients. Arch Orthop Trauma Surg 1986;105:3378. Eskola A, Vainionpaa S, Myllynen P, Patiala H, Rokkanen P. Surgery for ununited clavicular fracture. Acta Orthop Scand 1986;57:366-7. Joukainen J, Karaharju E. Pseudarthrosis of the clavicle. Acta Orthop Scand 1977;48:550-1. Jupiter JB, Leffert RD. Non-union of the clavicle: associated complications and surgical management. J Bone Joint Surg Am 1987;69A:753-60. Karahar u E, Joukainen J, Peltonen J. Treatment of pseudarthrosis of the c avic e. Injury 1982; 13:400-3. Manske DJ, Szabo RM The operative treatment of mid-shaft clavicular non-unions. J Bone Joint Surg Am 1985;67A: 1367-71. Muflai AB, Jupiter J8. Low-control dynamic compression pating of the clavicle. Injury 1994;25:41-5. Neer CS. Fractures of the distal third of the clavicle. Clin Orthop 1968;58:43-50. Neer CS. Nonunion of the clavicle. JAMA 1960;172: 1006-11. Nordqvist A, Petersson C, Redlund Johnell I. The natural course of lateral clavicle fracture. 15 (11-21) year follow-up of 110 cases. Acta Orthop Scand 1993;64:87-91. PyperJB. Non-union of fractures of the clavicle. Injury 1978; 9:268-70. Rabenseifner L. Etiology and therapy of clavicular-pseudarthrosis [author's translation]. Aktuel Traumatol 1981 ; 11 : 130-2. Sakellarides H. Pseudarthrosis of the clavicle. J Bone Joint Surg Am 1961 ;43A: 130-8. Wilkins RM, Johnston RM Ununited fractures of the clavicle. J Bone Joint Surg Am 1983;65A:773-8.
1-800-55-MOSBY This number links you to the full text of articles published in more than 25,000 journals, including all Mosby journals. MOSBY Document Express| a rapid response information retrieval service, provides quick turnaround, 24-hour availability, and speedy delivery methods. For inquiries and pricing information, call our toll-free 24-hour order line: 1-800-55-MOSBY; outside the United States: 415-259-5046; fax: 415r259-5019; E-mail: [email protected]. MOSBY Document Express~ is offered in cooperation with Dynamic Information Corp.
We studied the results of 16 consecutive midshaft clavicular nonunions operated on at the Shoulder and Elbow Clinic during the period from 1990 to 1993. All patients were treated with rigid 3.5 mm plate fixation and autologous cancellous bone grafting. Union of the fractures was achieved in all except one case, with a reconstruction ratio (restoration of bone length) of 0.96 (range 0.88 to 1.03). At follow-up 12 of 16 patients had returned to their preinjury activity level and according to the Constant score had obtained an excellent result. Two patients were graded as good, one as fair, and one had a failure. Thirteen of 16 patients were satisfied with the cosmetic outcome, assessing their cosmetic result as either good or excellent. Rigid plate fixation and restoration of clavicular length with autologous cancellous bone graft is recommended for the treatment of symptomatic clavicular midshaft nonunions. (J SHOULDERELBOWSURG 1995;4:337-44.)
Fracture of the clavicle is common, comprising between 5% and 16% of all fractures seen in a casualty department.* The incidence of nonunions is reported to be between O. 1% and 5%. 2' 6-8, 12, 15-17 Most nonunions and fractures appear in the middle one third of the clavicle, s' z. 15 A clavicular nonunion is usually diagnosed by lack of fracture healing 4 to 6 months after injury.S, 12, 17, 19, 20 It is generally agreed that surgery is indicated only in symptomatic nonunions.'~ ,6, 19, 2o Previous reports describe moderate to severe pain, grating, cosmetic deformity, and weakness of the shoulder in most of the symptomatic cases. 5" 8, lO, 12, 15, 18-2oVarious surgical and nonsurgical techniques have been described for the treatment of this potentially disabling condition.2, ~-8, lO, 12, 17-2oExternal or internal fixation of a clavicular nonunion is demanding because *References 2, 7, 8, 12, 15, 16, 18. From the Shoutder and Elbow Clinic, Department of Orthopaedic Surgery, Universib, Hospital of Aarhus. Reprint requests: Bo Sanderhoff Olsen, MD, Orthopaedic Research Fellow, Biomechanics Laboratory and Shoulder and Elbow Clinic, Universib, Hospital of Aarhus, Randersvej 1, DK-8200 Aarhus N, Denmark. Copyright @ 1995 by Journal of Shoulder and Elbow Surgery Board of Trustees. 1058-2746/95/$5.00 + 0 3 2 1 1 / 6 7 6 6 9
of the basic anatomy and biomechanics of the bone. s' 1o During recent years several authors have reported good results after surgical treatment of clavicular nonunions by using plate fixation and autologous iliac crest bone grafting.2, s, 6, 8-,3, 17, 18, 2o However, most studies include only a limited number of cases treated over a period of several years. 2" 5, 6, 8-1s, lZ, 18, 2o This treatment is claimed to be successful with a high ratio of fracture healing, but only two reports evaluate the functional and cosmetic outcome of the surgery. 1~ 12 In this retrospective study we reviewed the functional and cosmetic outcome of 16 consecutive patients with symptomatic clavicular midshaft nonunions treated with rigid internal fixation and autologous cancellous bone grafting.
PATIENTS AND METHODS Seventeen consecutive patients with a clavicular midshafl nonunion were treated at the Shoulder and Elbow Clinic, Aarhus University Hospital, during the period from early 1990 to late 1993. The patient data are presented in Table I. One patient was lost to follow-up and was excluded from the study, leaving 16 patients for study. Ten patients were men, and six were women. Twelve fractures were right sided and four left sided. The dominant 337
338
J. Shoulder Elbow Surg. September/October 1995
Olsen, Vaesel, and Sojbjerg
Table I Data for all patients in study Patient No.
Age at operation (yr)/sex
Injury/dominant side
Mechanism of injury
Duration of nonunion (wk)
Initial treatment
1
35/M
R/R
Car accident
2
35/F
R/R
Sport/fall
3
30/M
R/R
Bicycle accident
20
Figure of eight 14day delay
4
18/F
L/R
Car accident
18
Plating
5
49/M
R/R
Motorcycle accident
33
Arm sling 3 days
6
55/F
R/R
Bicycle accident
44
Figure of eight < 14 days
7
49/M
R/R
Bicycle accident
60
Figure of eight
8
32/M
R/R
Car accident
9
25/M
R/L
Bicycle accident
26
Figure of eight
10
38/F
L/R
Bicycle accident
25
Figure of eight 3-day delay
11
48/M
L/R
Motorcycle accident
22
Figure of eight
12
19/F
R/R
Car accident
50
!3
31/M
R/R
Motorcycle accident
90
Arm sling 14-day delay Arm sling 10 wk
14
43/M
R/R
Simple fall
65
Arm sling <14 days
15
13/F
L/R
Simple fall
164
Figure of eight 7 wk
16
35/M
R/R
Car accident
988
Figure of eight
16
Armsling <1 wk
144
Figure of eight
390
Arm sling < 14 days
Excellent,Excellent result, Radiographic union; z~CS(Constant scoring) <10. Activity normalized; Good, Good result: Radiographic union. ACS <20. AActivity _<2;Fair,Fair result: Radiographic union. ACS <30. z~J~ctivity_<4;Failed,Failed result: No radiographic union.
arm was affected in 11 patients. In three patients the fracture was sustained after a simple fall, in five it was the result of a bicycle accident, and in nine it was caused by a high-energy traffic accident. All fractures were classified according to Allman 1 and Neer TM as belonging to group 1 (midshaft) clavicular fractures. All fractures were initially treated conservatively with a figure-of-eight bandage or an arm sling for 3 to 61 days, except for one fracture primarily treated elsewhere with a compression plate. The time from trauma to final surgery was a mean of 135 weeks (range 16 to 988 weeks). The mean age of the pa-
tients at surgery was 34 years (range 13 to 55 years). Nine patients had an atrophic nonunion and six a hypertrophic nonunion. One patient had a failed osteosynthesis. On gross examination nine of the nonunions were displaced, presenting an overlap, diastasis, or both equal to or greater than the diameter of the clavicle (Fig. 1, A). Four nonunions presented with angulation exceeding 25 ~ The indication for surgery in 14 of 16 patients was pain in the affected shoulder during exercise or at rest. Seven patients had moderate pain only present at exercise, and seven described the pain as severe with pain at rest. Two patients were free
J. Shoulder Elbow Surg. Volume 4, Number 5
Fracture type and overlap
Olsen, Vaesel, and Scsjbjerg
Length of Postoperative clavicular Outcome follow-up (mo) length ratio of operation
Hypertrophic
51
0.97
Excellent
Hypertrophic > 1 bd overlap Hypertrophic Angulation 30 ~ Failed plating
30
0.92
Excellent
30
0.98
Excellent
34
0.93
Fair
Atrophic >2 bd diastasis Atrophic >3 bd overlap
30
0.92
Excellent
39
0.96
Good
Atrophic >2 bd overlap + diastasis Atrophic
21
0.99
Excellent
24
1.01
Excellent
Atrophic >2 bd diastasis Hypertrophic Angulation 30 ~ Atrophic >4 bd overlap Atrophic >4 bd overlap Atrophic >3 bd overlap Hypertrophic Angulation 25 ~ Hypertrophic Angulation 60 ~
18
1.00
Excellent
29
0.96
Excellent
28
1.03
Excellent
12
1.00
Fair
12
0.98
Excellent
12
0.96
Excellent
12
0.88
Excelbnt
14
0.90
Failed
Atrophic >3 bd overlap and diastasis
of pain but had a grating sensation during heavy work. In these two patients the primary indication for surgery was cosmetic. At operation the patient was placed in the semiseated position with the head turned to the opposite shoulder. The anterior iliac crest was prepared to obtain the bone graft. All patients received prophylactic antibiotics. A horizontal skin incision along the caudal edge of the clavicle was used. The normal clavicle and as much as possible of the nonunion was~ exposed subperiosteally. In the hypertrophic nonunions the hypertrophic callus was reduced to normal size, and interposed tissue was excised. In the atrophic cases the sclerotic ends were resected, and in both types the medullary canal was opened and reamed with a drill.
339
Comments Normal pain free ROM Normal pain free ROM Normal pain free ROM Normal pain free ROM. Sequelae from pelvic fracture Normal pain free ROM Pain in internal and external rotation Decreasedinternal rotation Pain free ROM Normal pain free ROM Normal pain free ROM Normal pain free ROM Normal pain free ROM Abduction 150 ~ Pain during ROM Normal pain free ROM Normal pain free ROM Prior ipsilateral fracture Pain free normal ROM Painful decreased ROM
After the nonunion was released, the shoulder tended to slide posteriorly, thereby restoring the length of the clavicle. A 3.5 mm plate was contoured to shape the clavicle exactly and was temporarily fixed to the clavicle by two bone clamps. The distance from the acromial tip to the midsterhum was palpated bilaterally to approximate the correct length of the clavicle, and the plate was secured to the bone with a minimum of six screws (Fig. 2). Autologous cancellous bone was harvested from the anterior iliac crest. The bone graft was packed into the clavicular defect, and in two cases with cortical grafting the intercalcary segment was fixed with a screw. The trapezius and deltoid fascia were carefully repaired. An arm sling was used for postoperative immo-
340
J. Shoulder Elbow Surg. September~October 1995
Olsen, Vaesel, and S~jbjerg
Table II Cosmetic score Cosmetic outcome
Preoperative
Postoperative
Excellent Good Acceptable Poor Bad
0 0 1 7 8
3 10 1 1 1
bilization for at least 21 days. Only passive motion of the shoulder not exceeding 90 ~ of elevation was allowed for the first 6 weeks. Heavy work was allowed after 3 months, and in general contact sports were allowed after 6 months. In 13 patients removal of the osteosynthesis material was necessary for pain or cosmetic reasons at an average of 11.6 months (range 3.5 to 27 months) after the operation. The follow-up time was a mean of 24 months (range 12 to 51 months). At follow-up, x-ray films, including those of the complete shoulder girdle, which allowed comparison of the clavicular length on the operated and nonoperated side, were obtained. This x-ray examination enabled evaluation of the restoration of clavicular length, which was calculated as the ratio between clavicular length on the injured and the noninjured side and the exact difference in millimeters. The cosmetic results were assessed by the patients and graded on a 1 to 5 scale, 1 being excellent and 5 being bad (Table II). In all the patients the activity scores were recorded for the time before the injury, after the trauma but before the operation, and after the operation. Patients were graded on a 1 to 8 scale, 1 expressing the highest activity level and 8 the lowest (Table III). Finally, the Constant shoulder score for assessing shoulder function was applied. 3' 4 This scoring allows evaluation of differences between injured and noninjured shoulders, recording pain, range of movement, and the objective force during combined abduction and elevation; these factors are measured with the Isobex traction device (Isobex, Curmed, Zurich Switzerland) (Table IV). The results were evaluated with a Student t test, paired and unpaired (SOLO; BMDP statistical software, Los Angeles, Calif.) and were considered significant at p < 0.05. RESULTS Solid healing was achieved in 15 of 16 patients (Figure 1, C). After the operation a minor clavicu-
lar shortening on the injured side was observed after x-ray evaluation. An average shortening of 6 mm (range -4 to 16 mm) was measured, with a mean clavicular reconstruction ratio of 0.96 (range 0.88 to 1.03) (Table I). Before the operation several patients had a marked shortening of their shoulder girdle as a result of the nonunion, with an overlap of the fracture ends of up to four times the bone diameter of the clavicle. This approximates a shortening of up to 80 mm (Table I and Figure 1, A, B). All the patients had forward and downward dropping of the injured shoulder compared with the noninjured side before the operation. This condition was usually described as hanging and shortening of the shoulder. Some patients also observed the medial fragment pointing proximally, threatening the integrity of the skin. Surgery cured the reported dropping in 15 of 16 patients, but in two of these patients a hump was observed at the former fracture site, reducing the cosmetic outcome. One patient (case 16) had an unchanged cosmetic score (Table II). After the operation a median gain of two classes on the cosmetic scale was reported compared with the preoperative observations (p < 0.001, paired t test). Thirteen of the 16 patients were satisfied with the cosmetic outcome of surgery, grading their result as either good or excellent (Table II). Fourteen of 16 patients had pain and functional disability before operation. After the operation 11 patients were free of pain, three had a mild grating sensation in the shoulder during exercise, and two patients had moderate pain, with pain being present only during heavy work. No cases of severe or disabling pain were seen after the ope r a t i o n . Twelve of 16 patients returned to their pretraumatic activity level after surgery (Table III); these were all rated as having an excellent surgical outcome (Table I). Three of the remaining four patients (cases 4, 6, 12) had some improvement but at the time of follow-up had not been able to resume their pretraumatic activity level, although healing of the clavicular nonunion occurred in all these patients. In one patient (case 16) no union was achieved. This patient had the longest period from trauma to final surgery (19 years). He had previously been treated with resection of a part of the medial fragment. A cortical graft from the iliac crest was used during the plating procedure to span the gap between the bone ends. Healing of this graft was
J. Shoulder Elbow Surg. Volume 4, Number 5
Olsen, Veesel, and S~jbjerg
341
Figure 1 Radiograph of patient 6. A, Preoperativex-ray film showing large overlap of atrophic nonunion. B, Postoperativex-ray film. C, Postoperativex-ray film showing operative outcome after removal of plate. never achieved. The compression plate had to be removed because of loosening and protrusion of the screws, threatening the skin. At follow-up this patient had a reduced and painful range of mo-
tion. He is now claiming workman's compensation. A median gain of three classes in the activity scale was reported after the operation compared
342
J. Shoulder Elbow Surg. September~October 1995
Olsen, Vaesel, and S~jbjerg
j Figure 2 Reconstruction of clavicular nonunion with plate and autologous
cancellous bone graft in gap.
Table III Activity score Activity level 1. 2. 3. 4. 5. 6. 7. 8.
ProFessionalsport (shoulder-demanding) NonproFessional sport (shoulder-demanding) Professionalsport (non-shoulder-demanding) Shoulder demanding profession Nonprofessional sport (non-shoulder-demanding) Non-shoulder-demanding profession Unemployed, not because of shou(der Unemployed, because of shoulder
with the posttraumatic activity level (p < 0.001, paired ttest). Evaluation of the pretraumatic activity compared with the postoperative activity showed a median decrease of one class (p < 0.05, paired t test). The Constant score of the injured and the noninjured shoulders at follow-up are listed in Table IV. These scores show a significantly reduced score for the injured shoulder compared with that of the noninjured shoulder (p < 0.02, paired t test). Patients operated later than 40 weeks after the initial injury had a significantly poorer outcome (p < 0.02, t test) than the patients operated in an earlier phase. Furthermore the atrophic nonunions had a poorer outcome compared with the hypertrophic nonunions (p < 0.03, t test). Apart from this finding we found no correlation between the outcome of surgery evaluated through the Constant score compared with age, sex, preoperative or postoperative clavicular shortening, and dominant or
Pretraumatic
Posttraumatic
Postoperative
0 6 0 8 2 0 0 0
0 1 0 0 2 5 0 8
0 3 0 8 4 1 0 0
nondominant arm. No neurologic dysfunction or thoracic outlet signs were observed after the operation. In this study no cases of superficial or deep infections were observed. Except for patient 16, who had loosening of the screws, no complications developed. DISCUSSION The clavicle provides stability and force to the shoulder joint by serving as a bony stabilizer and as a link between the thorax and the shoulder girdle, s This explains the described functional and cosmetic consequences of nonunion after a clavicular fracture, which were also observed in this study, lO, 15, 19 This study reports good functional outcome of plate fixation and autologous cancellous bone grafting in symptomatic, clavicular, midshafl nonunions, with grading of the functional outcome as excellent in 12 of 16 patients, and only one failure. This result is in accordance with previous
J. Shoulder Elbow Surg. Volume 4, Number 5
reports. 8, lO, 12 Furthermore this study exhibits good or excellent cosmetic results in 13 of 16 patients, with only two patients being dissatisfied with the final cosmetic outcome. In this study all but one patient returned to their pretraumatic occupations. This result corresponds to that in the report from one other author. 8 Several factors have been reported predisposing to nonunion after a clavicular fracture: (1) inadequate initial immobilization after closed or open reduction, s' lo, 13, 19 (2) severity of fracture displacement/' 8, lo, is, 19, 2o (3) primary open reduction of the fracture, s' lO, is, 19 (4) age (children rarely present with posttraumatic clavicular nonunions).S In this series 8 of 16 patients were noncompliant with the initial immobilization of 21 days. Some of these patients had an initial delay before immobilization of several days; others immobilized the shoulder girdle only during a limited period of less than 14 days (Table I). Only 7 of 16 cases underwent initial immobilization for at least 21 days. This finding supports the theory that inadequate immobilization of clavicular fractures favors the development of nonunions, s" lo, 13, 14 CraigS stated that fractures of the clavicle should be immobilized until healing, reported as being 4 to 6 weeks in young adults and 6 to 8 weeks in older patients. Nine of 16 patients initially had severe grades of fracture displacement presenting as overlap of more than one times the bone diameter. Furthermore four patients had angulation exceeding 25 ~. This finding confirms reports stating that the degree of displacement is an important predictor of the development of nonunion after a clavicular fracture.S, lo, is, 2o This study included only one child who was 13 years old. When this rate is compared with the rather high incidence of clavicular fractures among children, it tends to support the theory that children rarely have clavicular nonunions, s In this study the Constant score was significantly lower in the affected shoulder compared with that in the nonaffected shoulder, although the dominant arm was more frequently affected than the nondominant arm. Other authors found no association between the injured side and hand dominance of the patient. 1~ 2o A long period of conservative treatment of a symptomatic clavicular nonunion is recommended by some authors, who state that atrophic nonunions become less symptomatic with time and
Olsen, Vaesel, and Sojbjerg
343
Table IV Constant score at follow-up Patient No.
Injured
Noninjured
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
97 88 99 81 100 73 87 92 100 89 90 72 97 98 79 65
98 85 100 88 100 88 89 100 100 90 97 100 100 100 82 100
therefore do not necessarily require surgery. 2~ This belief was not supported by this study, because all atrophic nonunions were observed to be symptomatic in spite of a long follow-up. Furthermore the two patients with only cosmetic symptoms had hypertrophic nonunions.12 In this study the atrophic nonunions had a significantly poorer outcome than did the hypertrophic cases when compared according to Constant score and activity level. The only failure in this study was initially an atrophic nonunion. These results support the opinion of Connolly and Dehne 2 that the atrophic nonunions have a poor prognosis. This study also indicates that early operation of clavicular nonunions gives a better functional outcome.
In two cases free cortical lilac grafts were used to span clavicular defects. One of these cases failed. Other authors using free tibial or lilac grafts reported failure rates of 50%. is Different authors have also reported subjective weakness in abduction of the affected shoulder. Whether the weakness is caused by pain or by clavicular shortening is not specified. 8' 14, 2o No significant objective differences in force transmission over the affected shoulder girdle compared with that over the normal side were observed in this study. Force transmission was measured as part of the Constant scorings. The patients in this study all underwent careful reconstruction. This reconstruction yielded small differences in length between the operated and normal sides (Table I). On the basis of this series we conclude that the
344
J. Shoulder Elbow Surg.
Olsen, Vaesel, and S~jbjerg
September~October
restoration of clavicular length is essential to obtain good functional and cosmetic results. This study indicates that for a clavicular midshaft nonunion, the present operative technique offers a safe surgical method. This method has a predictably good end result concerning healing of the nonunion, which was achieved in 94% of the cases, and a good functional and cosmetic outcome for the patient. The results from this study and earlier reports suggest that rigid plating with autologous cancellous bone grafting is the treatment of choice for the patient with a symptomatic clavicular midshafl nonunion.* REFERENCES
7. 8. 9. 10. 11. 12. 13.
1. AIIman FLJ. Fractures and ligamentous injuries of the clavicle and its articulation. J Bone Joint Surg Am 1967;49A:774-84. 2. Connolly JF, Dehne R. Nonunion of the clavicle and thoracic outlet syndrome. J Trauma 1989;29:112732. 3. Constant CR. Assessment of shoulder function. Orthopade 1991 ;20:289-94. 4. Constant CR, Murley AH. A clinical method of functional assessment of the shoulder. Clin Orthop 1987; 160-4. 5. Craig ED. Fractures o~ the clavicle. In: Rockwood CA, Matsen FA, editors. The shoulder. Philadelphia: Saunders, 1990:367-412. 6. Edvardsen P, Odegard O. Treatment oF posttraumatic *References 6, 8, 10, 12, 13, 18, 20.
14. 15. 16. 17. 18. 19. 20.
1995
clavicular pseudarthrosis. Acta Orthop Scand 1977;48: 456-7. Eskola A, Vainionpaa S, Myllynen P, Patiala H, Rokkanen P. Outcome of clavicular fracture in 89 patients. Arch Orthop Trauma Surg 1986;105:3378. Eskola A, Vainionpaa S, Myllynen P, Patiala H, Rokkanen P. Surgery for ununited clavicular fracture. Acta Orthop Scand 1986;57:366-7. Joukainen J, Karaharju E. Pseudarthrosis of the clavicle. Acta Orthop Scand 1977;48:550-1. Jupiter JB, Leffert RD. Non-union of the clavicle: associated complications and surgical management. J Bone Joint Surg Am 1987;69A:753-60. Karahar u E, Joukainen J, Peltonen J. Treatment of pseudarthrosis of the c avic e. Injury 1982; 13:400-3. Manske DJ, Szabo RM The operative treatment of mid-shaft clavicular non-unions. J Bone Joint Surg Am 1985;67A: 1367-71. Muflai AB, Jupiter J8. Low-control dynamic compression pating of the clavicle. Injury 1994;25:41-5. Neer CS. Fractures of the distal third of the clavicle. Clin Orthop 1968;58:43-50. Neer CS. Nonunion of the clavicle. JAMA 1960;172: 1006-11. Nordqvist A, Petersson C, Redlund Johnell I. The natural course of lateral clavicle fracture. 15 (11-21) year follow-up of 110 cases. Acta Orthop Scand 1993;64:87-91. PyperJB. Non-union of fractures of the clavicle. Injury 1978; 9:268-70. Rabenseifner L. Etiology and therapy of clavicular-pseudarthrosis [author's translation]. Aktuel Traumatol 1981 ; 11 : 130-2. Sakellarides H. Pseudarthrosis of the clavicle. J Bone Joint Surg Am 1961 ;43A: 130-8. Wilkins RM, Johnston RM Ununited fractures of the clavicle. J Bone Joint Surg Am 1983;65A:773-8.
1-800-55-MOSBY This number links you to the full text of articles published in more than 25,000 journals, including all Mosby journals. MOSBY Document Express| a rapid response information retrieval service, provides quick turnaround, 24-hour availability, and speedy delivery methods. For inquiries and pricing information, call our toll-free 24-hour order line: 1-800-55-MOSBY; outside the United States: 415-259-5046; fax: 415r259-5019; E-mail: [email protected]. MOSBY Document Express~ is offered in cooperation with Dynamic Information Corp.